**A First Approach to Assess the Impact of Bottom Trawling Over Vulnerable Marine Ecosystems on the High Seas of the Southwest Atlantic**

J. Portela, J. Cristobo, P. Ríos, J. Acosta, S. Parra,

J.L. del Río, E. Tel, V. Polonio, A. Muñoz,

T. Patrocinio, R. Vilela, M. Barba and P. Marín

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/59268

**1. Introduction**

The Southwest Atlantic (SW Atlantic), corresponding to FAO Statistical Area 41, includes a total continental shelf area of approximately 1.96 million km2 of which a large portion lies off the Argentine coast (the Patagonian Shelf) and extends beyond Exclusive Economic Zones (EEZs) in the region [1-3]. This area is therefore integrated in the Southeast South American Shelf Large Marine Ecosystem (SSASLME) [4,5]. Currently, this region is the only worldwide significant area for high seas (HS) fisheries not covered by any Regional Fisheries Management Organisation (RFMO) [3].

The Patagonian Shelf (PS) hosts some of the most important fisheries in the world, targeting cephalopods (*Illex argentinus* [Castellanos, 1960] and *Doryteuthis gahi* [D'Orbigny, 1835]), and hakes (*Merluccius hubbsi* [Marini, 1933] *Merluccius australis* [Hutton, 1872]) [3,6-14]. Most of the exploited demersal stocks on the HS are straddling stocks, including Argentine shortfin squid (*I. argentinus*), Argentine hake (*M. hubbsi*) and southern blue whiting (*Micromesistius australis* [Norman, 1937]) [15].

Several authors [2,3,16-23] have studied the potential disturbance of the seabed by bottom otter trawls and the possible negative effects on the structure of benthic communities. In recent years, several resolutions of the United Nations General Assembly [24-28] on sustainable fisheries made a call to States and RFMOs to identify vulnerable marine ecosystems (VMEs)

© 2015 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and eproduction in any medium, provided the original work is properly cited.

and determine whether bottom fishing activities would cause a significant adverse impact on such ecosystems.

Sensitive species such as deep-water corals and deep-water sponges are found throughout the world oceans. Thus, the importance of habitat-structuring organisms is not restricted to shallow water, but also to shelf-break, hydrothermal vents, seamounts, and even the once considered constant and uniform deep-sea basins. Deep-water corals are vulnerable organisms occurring in the upper bathyal zones throughout the world and threatened by human activities, particularly fishing and oil exploration [29-31]. Fishing has a significant adverse impact (SAI) on deep-water coral communities in all oceans [32-35], particularly in the Northeast and Northwest Atlantic [36-40], Northeast Pacific [41,42], and Southwest Pacific [43-46]. In the SW Atlantic, the HS are one of the areas where deep-sea science has, to date, not been very active.

Protection of VMEs is a significant element of the management framework for bottom fisheries in high seas areas of the world ocean and its identification for selecting suitable protection areas is a challenge that conventional fisheries science cannot alone solve satisfactorily. Instead, it requires a multidisciplinary approach [21,22,47]. From the point of view of management of bottom fisheries and the governance of high seas areas, the situation in the PS poses an added problem as there is no any RFMO in force [2]. In its 2014 report [48], the Global Ocean Commission (GOC) recognises that continued scientific research is necessary to assess the cumulative impacts of human activities on the high seas so that informed decisions can be made about reversing the degradation of the global ocean.

Submarine canyons are unique habitats in terms of complexity, instability, material processing, and hydrodynamics. They may support diverse assemblages of larger epibenthos [49]. Inside canyons, abundance and diversity of the macrofauna depend, to some extent, on the physical disturbance regime and on the rate and quantity of organic matter deposited. In the study area, canyons and submarine mounts were shown to be hot spots of benthic biodiversity of species and ecosystems.

Benthos refers to the community of organisms which live on, in, or near the seabed, also known as the benthic zone. Megabenthos or macrobenthos comprises the more visible, benthic organisms exceeding 1 mm in size and large enough to be determined on photographs [50,51]. Megabenthos is a key issue of environmental studies, as it represents a major fraction of the deep-sea benthic biomass and plays a key role in deep-sea ecosystems [52]. Tracey et al. (2007) in [53] reported linear and radial annual growth rates of 20 mm and 0.2 mm, respectively, for some genera of the ISIDIDAE Family (Lamouroux, 1812), which is presumably evidence of the high vulnerability of these taxa to direct or indirect mechanical impact produced by the sediment removal, re-suspension, etc. caused by bottom fishing activities.

Some of these organisms form complex 3D structures protruding from the seabed, allowing for the settlement of sessile species needing consolidated substrata to settle and develop (sponges, other cnidarians), and providing shelter and food for a wide range of vagile fauna (crustaceans, echinoderms, molluscs, and some fish).

## **2. Materials and methods**

and determine whether bottom fishing activities would cause a significant adverse impact on

Sensitive species such as deep-water corals and deep-water sponges are found throughout the world oceans. Thus, the importance of habitat-structuring organisms is not restricted to shallow water, but also to shelf-break, hydrothermal vents, seamounts, and even the once considered constant and uniform deep-sea basins. Deep-water corals are vulnerable organisms occurring in the upper bathyal zones throughout the world and threatened by human activities, particularly fishing and oil exploration [29-31]. Fishing has a significant adverse impact (SAI) on deep-water coral communities in all oceans [32-35], particularly in the Northeast and Northwest Atlantic [36-40], Northeast Pacific [41,42], and Southwest Pacific [43-46]. In the SW Atlantic, the HS are one of the areas where deep-sea science has, to date, not

Protection of VMEs is a significant element of the management framework for bottom fisheries in high seas areas of the world ocean and its identification for selecting suitable protection areas is a challenge that conventional fisheries science cannot alone solve satisfactorily. Instead, it requires a multidisciplinary approach [21,22,47]. From the point of view of management of bottom fisheries and the governance of high seas areas, the situation in the PS poses an added problem as there is no any RFMO in force [2]. In its 2014 report [48], the Global Ocean Commission (GOC) recognises that continued scientific research is necessary to assess the cumulative impacts of human activities on the high seas so that informed decisions can be

Submarine canyons are unique habitats in terms of complexity, instability, material processing, and hydrodynamics. They may support diverse assemblages of larger epibenthos [49]. Inside canyons, abundance and diversity of the macrofauna depend, to some extent, on the physical disturbance regime and on the rate and quantity of organic matter deposited. In the study area, canyons and submarine mounts were shown to be hot spots of benthic biodiversity of species

Benthos refers to the community of organisms which live on, in, or near the seabed, also known as the benthic zone. Megabenthos or macrobenthos comprises the more visible, benthic organisms exceeding 1 mm in size and large enough to be determined on photographs [50,51]. Megabenthos is a key issue of environmental studies, as it represents a major fraction of the deep-sea benthic biomass and plays a key role in deep-sea ecosystems [52]. Tracey et al. (2007) in [53] reported linear and radial annual growth rates of 20 mm and 0.2 mm, respectively, for some genera of the ISIDIDAE Family (Lamouroux, 1812), which is presumably evidence of the high vulnerability of these taxa to direct or indirect mechanical impact produced by the

Some of these organisms form complex 3D structures protruding from the seabed, allowing for the settlement of sessile species needing consolidated substrata to settle and develop (sponges, other cnidarians), and providing shelter and food for a wide range of vagile fauna

sediment removal, re-suspension, etc. caused by bottom fishing activities.

(crustaceans, echinoderms, molluscs, and some fish).

made about reversing the degradation of the global ocean.

such ecosystems.

2722 Biodiversity in Ecosystems - Linking Structure and Function

been very active.

and ecosystems.

In accordance with the aforementioned UNGA resolutions [24-28] and the FAO deepwater guidelines [54], the Spanish Institute of Oceanography (Instituto Español de Oceanografía [IEO]) conducted from October 2007 to April 2010 a series of 13 multidisciplinary research cruises on the HS of the SW Atlantic, to identify VMEs and to assess the potential interactions with fishing activities. This paper presents the results of the five first cruises, consistently with UNGA resolutions (paragraphs 80 and 83 to 87 of resolution 61/105 (2007) and paragraphs 117 and 119 to 127 of resolution 64/72 (2010) in [27,28], which support making publicly available information on interactions between bottom fisheries and VMEs in the HS.

The use of spatial management tools to preserve the marine biodiversity of species inhabiting the HS has been broadly discussed in recent years [55]. To make such spatial management possible, our immediate objectives are: assessing specific biodiversity (mainly describing new species to science); describing the different habitats, ecosystems and deep-sea geomorpholog‐ ical features identified; and analysing their interactions and relationships to protect the full range of potentially different habitats.

The explored area during the five cruises conducted between October 2007 and April 2008 (Table 1) was located on the southern part of the HS of the SW Atlantic, to the east of the Argentinian EEZ 200 miles limit and between 44° 40'S and 47° 51'S up to the 1500 m depth contour (Figure 1). The rest of the study area (up to 42°S) was surveyed during the eight following cruises (October 2008-April 2010), but the analysis of the information concerning VMEs collected during those last cruises, is still ongoing.


**Table 1.** Cruises carried out by R/V "Miguel Oliver".

In the right image of Figure 1 a non coloured area in the shelf can be roughly appreciated around 45°30'S and between 60°00'W-60°40'W, for which it was not possible to collect multibeam bathymetry data (no data) due to bad sea state conditions. The exploration of this area was carried out during one of the cruises conducted in 2009. Nevertheless, this type of data is not relevant for the present study, for which several trawl and CTD stations allowed the collection of pertinent information. The blue lines in the left image of Figure 1 correspond‐ ing to the 600, 1000 and 1500 m depth contours.

Key concepts for definition of VMEs were applied according to the FAO International Guidelines for the Management of Deep-Sea Fisheries in the High Seas [54]. These guidelines classify marine ecosystems as vulnerable based on several criteria: (1) uniqueness or rarity; (2) functional significance of the habitat; (3) fragility; (4) life-history traits of component species that make recovery difficult; and (5) structural complexity.

**Figure 1.** Study area and positioning of the stations carried out during the research cruises onboard the R/V "Miguel Oliver".

For an adequate identification of VMEs, the two approaches in operation since 2008 by the NAFO Scientific Committee and the NAFO Working Group on Ecosystem Approach to Fisheries Management (WGEAFM) were applied in this study [56,57]: (1) the examination of cumulative catch data by ranking the biomass of VME taxa in each trawl from lowest to highest and then plotting the increase in cumulative biomass with each additional trawl; and (2) the use of Geographical Information System (GIS) to map the density of vulnerable species and groups' by-catch [58].

Key concepts for definition of VMEs were applied according to the FAO International Guidelines for the Management of Deep-Sea Fisheries in the High Seas [54]. These guidelines classify marine ecosystems as vulnerable based on several criteria: (1) uniqueness or rarity; (2) functional significance of the habitat; (3) fragility; (4) life-history traits of component species

**Figure 1.** Study area and positioning of the stations carried out during the research cruises onboard the R/V "Miguel

For an adequate identification of VMEs, the two approaches in operation since 2008 by the NAFO Scientific Committee and the NAFO Working Group on Ecosystem Approach to

Oliver".

that make recovery difficult; and (5) structural complexity.

2744 Biodiversity in Ecosystems - Linking Structure and Function

The study area included part of the outer shelf and upper and middle slope of the PS and was divided into thirteen depth strata in order to obtain a higher resolution in the description of vulnerable organisms. The research cruises involved five scientific disciplines: cartography, geology, benthos, fisheries, and hydrography.

This study used data from three main sources: i) Information from the five research cruises (geological, echosounder and oceanographic data; benthos and fish samples; fishery catch data [cpue]); ii) Data from commercial fishing activity collected by onboard scientific observers from 1989 to 2007 (fishery footprint); and iii) Commercial information on historical landings and effort data (provided by the Spanish fishing sector), as well as catch data for the main com‐ mercial species during the period 2000-2007 (logbooks filled in by captains of the fishing vessels, and provided by the Spanish General Secretariat for Fisheries [SGP]).

Geophysical and geological data were collected following internationally accepted standards and protocols for habitat mapping [20,59]. Full sea floor coverage using swath bathymetry provided a very high resolution of sea floor morphology. The backscatter data from multibeam echosounder together with high resolution seismic reflection profiles made available valuable data on the seabed sediments types. These data provided the geomorphological and acoustic basis to design a ground-truth planning strategy allowing for precise habitat mapping. Navigation during the surveys was via differential GPS Simrad GN33 using satellite correc‐ tions integrated into an inertial-aided Seapath 200 system. Swath-bathymetric data were acquired using a hull-mounted Kongsberg-Simrad EM 302 multibeam echosounder (288 individual beams, angular coverage up to 150°) operating at a frequency of 30 kHz. To correct the multibeam bathymetry, we carried out systematic casts of direct sound velocity profiles on the water column with an Applied Microsystems SV Plus equipment. Data processing included the removal of anomalies and the necessary sound velocity corrections using the Kongsberg–Simrad swath bathymetric software package NEPTUNE. Valid data were gridded at 50×50 m cell size resolution on a SUN workstation. The seismic parametric system Topas 18 produced very high resolution seismic profiles along all ship tracks. Sub-bottom penetration varied, according to the lithology, between 150 and 250 m. Morphometrical data were obtained using ArcGis (ESRI) and Fledermaus software (Interactive Visualization Systems [IVS]) to provide final 3D images of the seafloor morphology.

Samples of benthic fauna analysed in this study were collected with the Lofoten bottom trawl gear itself. Benthic fauna samples were sorted on board and preserved (70% ethanol or 4% buffered formaldehyde-seawater solution) for further identification analysis. Even if the bottom-trawl by-catch collected information did not allow for a detailed habitat mapping of VMEs, it provided a valuable indication of VME presence/absence that can be used to propose conservation measures, such as candidate areas for bottom fishery closures [23].

Sediment samples were collected using net collectors attached to the Lofoten fishing gear (Atlantis 2008 cruise) and with an USNEL type box-corer (BC) (maximum breakthrough of 60 cm; effective sampling area of 0.25 m2 [50×50 cm]). A few samples were taken using a Bouma type box-corer (effective sampling area of 0.0175 m2 [10×17.5 cm]). Both gears are designed to take undisturbed samples from the top of the seabed, and are suitable for almost every type of sediment. Sediment temperature and redox profiles (Eh) were immediately performed for the box-corer sample after each station. In the laboratory, the granulometrical analysis of the sediment was carried out by dry sorting the coarse fraction (>62 µm) and the sedimentation of the fine fraction (<62 µm). The organic matter content was assessed after calcinating (at 500°C for 24 h) and drying the sediment sample.

The hydrographical conditions in the studied area during the Atlantis 2008 cruise were characterised by means of a Seabird-25 CTD probe (SBE-25), equipped with oximeter, fluor‐ ometer and PAR detector. The survey schedule was optimized by systematically deploying the CTD at fishing stations below 500 m, but not always at greater depths. At each cast, the CTD was deployed to 5 m depth and stabilised for approximately 3 min. Once stable, the CTD was brought back to the surface and started profiling at a constant speed of 1 m⋅s-1. The SBE-25 worked in auto-contained mode at a frequency of 8 scans⋅s-1 and the downloaded data were converted into physical units and pre-processed by using the SeaBird software (SeaSave/SBE DataProcesing-win32) with standard calibration values. Quality control and post-processing was performed with MATLAB.

Atlantis 2008 stratified bottom trawl survey enabled the assessing of the biomass and bathy‐ metric distribution of the main commercial and most abundant fishery stocks by means of the swept area method. The survey used a stratified random design with strata boundaries definedby latitude and depth ranges, depth strata 1-7 located south of parallel 45°S and depth strata 8-13 sited north of the referred parallel (Table 2). Scheduled fishing stations (hauls of 30 min) were performed using a Lofoten bottom trawl net fitted with a rockhopper mix train with bobbins and rubber separators, suitable for deep-water fishing over irregular bottoms. Mean trawl speed was of 3.2 knots and trawl direction followed the bathymetric profile in the upper slope, but was variable in the outer shelf and middle slope.

Data recorded by scientific onboard observers from 1989 to 2007 between latitude 42°S and 48°S were used for mapping only the Spanish fishery footprint, since fishing data of other fleets were unavailable to us. The IEO observers' program placed one observer per selected vessel to cover 12% to 15% of the whole fleet. Table 3 summarize the activities (number of hauls year-1) of the IEO observers on the HS of the SW Atlantic, were Divisions 42 and 46 correspond to the areas roughly around parallels 42°S and 46°S.

Data used for each fishing haul corresponded to the middle tow position, since it offers more relevant information than the initial or final positions. All middle tow positions were imported into ArcGIS 9.3 mapping software to plot all trawl tows as straight lines between the reported start and end positions. They were then exported to a grid of 5'×10' min blocks, and any block including at least two tows was retained for mapping the bottom trawl footprint.

A First Approach to Assess the Impact of Bottom Trawling Over Vulnerable Marine Ecosystems on the High Seas… 7 http://dx.doi.org/ 10.5772/59268 277


**Table 2.** Scheme of hauls by depth stratum, and main characteristics (ATLANTIS 2008 cruise).

Sediment samples were collected using net collectors attached to the Lofoten fishing gear (Atlantis 2008 cruise) and with an USNEL type box-corer (BC) (maximum breakthrough of 60

type box-corer (effective sampling area of 0.0175 m2 [10×17.5 cm]). Both gears are designed to take undisturbed samples from the top of the seabed, and are suitable for almost every type of sediment. Sediment temperature and redox profiles (Eh) were immediately performed for the box-corer sample after each station. In the laboratory, the granulometrical analysis of the sediment was carried out by dry sorting the coarse fraction (>62 µm) and the sedimentation of the fine fraction (<62 µm). The organic matter content was assessed after calcinating (at

The hydrographical conditions in the studied area during the Atlantis 2008 cruise were characterised by means of a Seabird-25 CTD probe (SBE-25), equipped with oximeter, fluor‐ ometer and PAR detector. The survey schedule was optimized by systematically deploying the CTD at fishing stations below 500 m, but not always at greater depths. At each cast, the CTD was deployed to 5 m depth and stabilised for approximately 3 min. Once stable, the CTD was brought back to the surface and started profiling at a constant speed of 1 m⋅s-1. The SBE-25 worked in auto-contained mode at a frequency of 8 scans⋅s-1 and the downloaded data were converted into physical units and pre-processed by using the SeaBird software (SeaSave/SBE DataProcesing-win32) with standard calibration values. Quality control and post-processing

Atlantis 2008 stratified bottom trawl survey enabled the assessing of the biomass and bathy‐ metric distribution of the main commercial and most abundant fishery stocks by means of the swept area method. The survey used a stratified random design with strata boundaries definedby latitude and depth ranges, depth strata 1-7 located south of parallel 45°S and depth strata 8-13 sited north of the referred parallel (Table 2). Scheduled fishing stations (hauls of 30 min) were performed using a Lofoten bottom trawl net fitted with a rockhopper mix train with bobbins and rubber separators, suitable for deep-water fishing over irregular bottoms. Mean trawl speed was of 3.2 knots and trawl direction followed the bathymetric profile in the upper

Data recorded by scientific onboard observers from 1989 to 2007 between latitude 42°S and 48°S were used for mapping only the Spanish fishery footprint, since fishing data of other fleets were unavailable to us. The IEO observers' program placed one observer per selected vessel to cover 12% to 15% of the whole fleet. Table 3 summarize the activities (number of hauls year-1) of the IEO observers on the HS of the SW Atlantic, were Divisions 42 and 46 correspond

Data used for each fishing haul corresponded to the middle tow position, since it offers more relevant information than the initial or final positions. All middle tow positions were imported into ArcGIS 9.3 mapping software to plot all trawl tows as straight lines between the reported start and end positions. They were then exported to a grid of 5'×10' min blocks, and any block

including at least two tows was retained for mapping the bottom trawl footprint.

[50×50 cm]). A few samples were taken using a Bouma

cm; effective sampling area of 0.25 m2

2766 Biodiversity in Ecosystems - Linking Structure and Function

was performed with MATLAB.

500°C for 24 h) and drying the sediment sample.

slope, but was variable in the outer shelf and middle slope.

to the areas roughly around parallels 42°S and 46°S.


**Table 3.** Number of hauls/year and division recorded by scientific observers.

Proper identification of the areas where VMEs are present followed the methodology used by the NAFO in its Regulatory Area [60]. Threshold catches, defined as catch levels of significant concentrations of invertebrates to be considered as possible VME areas, were assessed by analysing the cumulative biomass frequencies. Cumulative catch curve method was chosen to calculate the threshold catch. The cumulative frequency was plotted for all capture sets where taxa, considered as vulnerable by the International Guidelines for the Management of Fisheries [54] and by the Convention for the Protection of the marine Environment of the North-East Atlantic (OSPAR), were identified. The threshold selection for each taxon was made on the basis of minimum/maximum catch, density and morphological characteristics. Once a location of significant concentrations of vulnerable organisms was defined (key location), a 2 nm radius buffer zone around it was drawn to provide a safe margin of error on site.

The Random Forest algorithm for classification (RF) was used to predict the potential distri‐ bution of vulnerable benthic species by rating environmental conditions on the basis of previous observations.

RF is a non-parametric statistical method for data analysis that makes no distributional assumptions about the predictor or response variables [61], showing high prediction accuracy classifying rocky benthic communities [62] and beating other methods commonly used for ecological prediction [61,63]; The algorithm calculate the suitability of a given habitat for a given species based on known affinities with habitat characteristics, stored as raster maps, and called independent ecogeographical variables (EGV). According to HSI values, a map of species' expected distribution is produced, a value ranging from 0 to 1 showing the probability that the habitat of a given location is suitable for the species occurrence [64]. Thus, for a particular location, high HSI values mean high chances of the species' occurrence. To perform this mapping, presence/absence data from different vulnerable benthic organisms found in the study area were used as dependent variables of the different EGV.

Gathering accurate sampling presence/absence data is a critical part of the study, since the absence of a species in a given location can be due to several reasons: the species is present but is not observed, the species is absent even though the habitat is suitable, or the species is absent because of the unsuitability of the habitat. Only the last reason is considered as a "true absence" [65,66]. As presence data were aggregated into one single group named "vulnerable organ‐ isms", the resulting HSI predicted the potential habitat of any of the considered vulnerable organisms in the HS of the SW Atlantic under study.

The RF method offers the possibility to calculate an accurate unbiased estimator, using Out-Of-Bag (OOB) observations as an internal validation data set [67], computed from the resulting confusion matrix [68]. Accuracy is the proportion of the total number of predictions that were correct and this accuracy indicator is offered to the user as a measure of the model's predictive performance. It is determined using the equation:

$$ACC\_{OOB} = \frac{\text{TS} + \text{TV}}{N}$$

Where TS is the number of truly suitable locations, i.e. suitable locations correctly classified by the model; TU is the number of truly unsuitable locations, in other words unsuitable locations that have been correctly classified; and N is the total number of observations.

Data was analyzed using the R statistical software [69] and the "Random Forest" package [70] and predictions were exported to a shapefile format using the "maptools" package [71]. GIS visualization of results was performed using the ESRI ArcMap 10.0 software.

Selected environmental variables involved in the study included depth, slope, sea bottom temperature, substrate characteristics and topographic position (Table 4). The topographic position was sorted into six categories: shelf (1), outcrop areas on the shelf (2), high slope (3), low slope (4), abyssal flats (5), and canyons (6).


**Table 4.** Summary of the environmental variables used for the Habitat Suitability Index (HSI) modelling of vulnerable organisms.

CTD stations' sea bottom temperature data were interpolated for the whole area using the local polynomial interpolation function (LPI) implemented in the ArcGIS 10.0 software. Slope was derived from the bathymetry high resolution data, and after studying the semivariogram, substrate characteristics were interpolated from granulometrical measures for the whole area using a universal kriging interpolator (Unpublished).

All the explanatory data were extracted for the presence/absence data locations, subsequently exported and analysed with the R statistical software using the BIOMOD package [72]. Several presence/absence models were performed: Generalized Additive Models [73,74], Multivariate Adaptative Regression Splines [74,75], Generalized Boosting Models [74,76] and Random Forest model (RF) [67,74].

## **3. Results**

Proper identification of the areas where VMEs are present followed the methodology used by the NAFO in its Regulatory Area [60]. Threshold catches, defined as catch levels of significant concentrations of invertebrates to be considered as possible VME areas, were assessed by analysing the cumulative biomass frequencies. Cumulative catch curve method was chosen to calculate the threshold catch. The cumulative frequency was plotted for all capture sets where taxa, considered as vulnerable by the International Guidelines for the Management of Fisheries [54] and by the Convention for the Protection of the marine Environment of the North-East Atlantic (OSPAR), were identified. The threshold selection for each taxon was made on the basis of minimum/maximum catch, density and morphological characteristics. Once a location of significant concentrations of vulnerable organisms was defined (key location), a 2 nm radius

The Random Forest algorithm for classification (RF) was used to predict the potential distri‐ bution of vulnerable benthic species by rating environmental conditions on the basis of

RF is a non-parametric statistical method for data analysis that makes no distributional assumptions about the predictor or response variables [61], showing high prediction accuracy classifying rocky benthic communities [62] and beating other methods commonly used for ecological prediction [61,63]; The algorithm calculate the suitability of a given habitat for a given species based on known affinities with habitat characteristics, stored as raster maps, and called independent ecogeographical variables (EGV). According to HSI values, a map of species' expected distribution is produced, a value ranging from 0 to 1 showing the probability that the habitat of a given location is suitable for the species occurrence [64]. Thus, for a particular location, high HSI values mean high chances of the species' occurrence. To perform this mapping, presence/absence data from different vulnerable benthic organisms found in the

Gathering accurate sampling presence/absence data is a critical part of the study, since the absence of a species in a given location can be due to several reasons: the species is present but is not observed, the species is absent even though the habitat is suitable, or the species is absent because of the unsuitability of the habitat. Only the last reason is considered as a "true absence" [65,66]. As presence data were aggregated into one single group named "vulnerable organ‐ isms", the resulting HSI predicted the potential habitat of any of the considered vulnerable

The RF method offers the possibility to calculate an accurate unbiased estimator, using Out-Of-Bag (OOB) observations as an internal validation data set [67], computed from the resulting confusion matrix [68]. Accuracy is the proportion of the total number of predictions that were correct and this accuracy indicator is offered to the user as a measure of the model's predictive

Where TS is the number of truly suitable locations, i.e. suitable locations correctly classified by the model; TU is the number of truly unsuitable locations, in other words unsuitable locations that have been correctly classified; and N is the total number of observations.

N

ACC OOB= TS + TU

buffer zone around it was drawn to provide a safe margin of error on site.

study area were used as dependent variables of the different EGV.

organisms in the HS of the SW Atlantic under study.

performance. It is determined using the equation:

previous observations.

2788 Biodiversity in Ecosystems - Linking Structure and Function

### **3.1. Geomorphology**

Geomorphological and geophysical data from the five research cruises revealed that the outer shelf was mantled by 15 m high sand ridges, and was 60 to 67 m deeper than the maximum 120 m lowering of sea-level during the last glacially induced regression. This difference in depth indicates that the PS had experienced subsidence in the Holocene. These ridges are relict and were probably constructed during the post-glacial transgression by the north flowing Falkland (Malvinas) Current, since they are resting on shell layers of <35,000 to 11,000 years old [77].

The upper continental slope descends from the shelf break, located at depths from 200 to 750 m, and is scarred by iceberg plough marks whose orientation and morphology suggest that icebergs carried northwards by the Falkland (Malvinas) Current were probably responsible for this erosion during the last glaciations [78].

Scattered over the study area (south of 45˚S) we found pockmarks, carbonate mounds formed by deep-water corals, northwards furrows, areas of smooth topography and sediment waves indicating that deposition on this part of the middle slope is controlled by bottom currents [79].

Seven submarine canyons were identified on the middle slope surveyed (Figure 2). Canyons 1 to 6 were cut by turbidity currents, whereas canyon 0 resulted from the combined effect of turbidity currents and coalescence of pockmarks (formed by the expulsion of thermogenic gas). These gas and fluid seepages contributed to the formation of canyons and to the partial detachment of blocks from the canyon walls. Thus, the thermogenic gas responsible for the formation of the identified pockmarks on the middle slope could be deep-seated, probably related to the Falkland Rift Basin, north of the Falkland (Malvinas) Islands [80,81].

**Figure 2.** Colour shaded 3D bathymetric map of a segment of the Patagonian Argentinian margin compiled from mul‐ tibeam backscatter data. Arabic numbers identify submarine canyons discussed in text. CS=Continental shelf; US=Up‐ per continental slope; MS=Middle continental slope; P=Pockmark; PL=Iceberg plough marks.

The association of gas seepage with deep-water corals has been reported by [82] in pockmarks off Brazil. If such association also occurs on the Patagonian margin, those communities may be quite widespread in our study area.

### **3.2. Benthic communities**

*Bathelia candida* (Moseley, 1881) was found to be one of the main reef builder species in the study area, providing habitat for diverse associated fauna of sponges, crustaceans, echino‐ derms, molluscs, and other cnidarians. The benthic megafauna caught during the cruises included invertebrates as well as Phyla Chordata and Hemichordata. Phyla Cnidaria and Porifera were dominant in terms of biomass (46% and 30%, respectively [Figure 3A]). The high abundance of Cnidaria is remarkable, since 33.7% of the biomass of this phylum corresponded to the Class Octocorallia, including significant groups such as gorgonians (sea fans), alcyona‐ ceans (leather corals) and pennatulaceans (sea pens). In addition, the VMEs dominated by suspensivore and/or filter feeding organisms are habitats with high biodiversity and many resources.

depth indicates that the PS had experienced subsidence in the Holocene. These ridges are relict and were probably constructed during the post-glacial transgression by the north flowing Falkland (Malvinas) Current, since they are resting on shell layers of <35,000 to 11,000 years

The upper continental slope descends from the shelf break, located at depths from 200 to 750 m, and is scarred by iceberg plough marks whose orientation and morphology suggest that icebergs carried northwards by the Falkland (Malvinas) Current were probably responsible

Scattered over the study area (south of 45˚S) we found pockmarks, carbonate mounds formed by deep-water corals, northwards furrows, areas of smooth topography and sediment waves indicating that deposition on this part of the middle slope is controlled by bottom currents [79].

Seven submarine canyons were identified on the middle slope surveyed (Figure 2). Canyons 1 to 6 were cut by turbidity currents, whereas canyon 0 resulted from the combined effect of turbidity currents and coalescence of pockmarks (formed by the expulsion of thermogenic gas). These gas and fluid seepages contributed to the formation of canyons and to the partial detachment of blocks from the canyon walls. Thus, the thermogenic gas responsible for the formation of the identified pockmarks on the middle slope could be deep-seated, probably

**Figure 2.** Colour shaded 3D bathymetric map of a segment of the Patagonian Argentinian margin compiled from mul‐ tibeam backscatter data. Arabic numbers identify submarine canyons discussed in text. CS=Continental shelf; US=Up‐

The association of gas seepage with deep-water corals has been reported by [82] in pockmarks off Brazil. If such association also occurs on the Patagonian margin, those communities may

*Bathelia candida* (Moseley, 1881) was found to be one of the main reef builder species in the study area, providing habitat for diverse associated fauna of sponges, crustaceans, echino‐ derms, molluscs, and other cnidarians. The benthic megafauna caught during the cruises

per continental slope; MS=Middle continental slope; P=Pockmark; PL=Iceberg plough marks.

be quite widespread in our study area.

**3.2. Benthic communities**

related to the Falkland Rift Basin, north of the Falkland (Malvinas) Islands [80,81].

old [77].

for this erosion during the last glaciations [78].

280 10 Biodiversity in Ecosystems - Linking Structure and Function

**Figure 3.** Biomass per Phyla in total strata (A) and by stratum < 200 (B), 201-300 (C), 301-400 (D), 401-1000 (E), 1001-1500 m depth (F).

A large part of the benthic samples contained erect sponges, octocorals, colonial scleractini‐ ans, calcified antipatharians, and hydrozoans (Family STYLASTERIDAE), all of them slowgrowing organisms considered as vulnerable by the UN and the OSPAR standards (see Table 7).

Bathymetric strata differences clearly arise by comparing the composition of the sampled benthic megafauna (Figures 3B-F):

Strata 1 and 8 (<200 m) showed a low catch of benthos (17,209 and 41,202 g. respectively), both in number and diversity. We observed a strong dominance of pectinid molluscs of the Genus *Zygochlamys* (Ihering, 1907) (60.39% of the biomass [Figure 3B]), mainly *Z. patagonica* (King & Broderip, 1832), followed by those of the Genus *Chlamys* (Röding, 1798). Vulnerable organisms were practically unrepresented in these shallower strata, probably due to the bottom trawling activities for years by bottom trawlers from international fleets.

Strata 2 and 9 (201-300 m) recorded the lowest catch in terms of biomass (2121 and 1576 g, respectively). In these strata, detritivorous and opportunistic species were predominant, and the presence of vulnerable organisms was negligible again. Compared to strata 1 and 8, we observed an increase of the benthic cnidarians' biomass values, dominated by gorgonians from Family PRIMNOIDAE (Milne Edwards, 1857) (Octocorallia; Gorgonacea) (80.32% in biomass, [Figure 3C]).

Strata 3 and 10 (301-400 m) were hardly sampled due to the reduced number of valid hauls (3 in stratum 3 and 2 in stratum 10). The low benthic biomass and the negligible presence of vulnerable organisms (Figure 3D) could be attributed to bottom fishing activities, as above‐ mentioned for strata 1 and 8.

Strata 4, 11, 5, 6, and 12 of intermediate depths (401-1000 m, [Figure 3E]) recorded high biomass and numbers of octocorals, sponges, colonial scleractinians (*Bathelia candida*), and large hydrocorals. Octocorals included colonies of various genera belonging to families PRIMNOI‐ DAE and ISIDIDAE. As aforementioned, the increase and proliferation of these species create complex 3D structures providing the ideal habitat for a wide range of organisms. In those strata, the large amount of filter feeders and suspensivore sessile organisms is an indication of the presence of unaltered, complex and structured ecosystems. In the future, ROV and other submersible camera systems could confirm these assumptions.

Strata 7 and 13 (1001-1500 m, [Figure 3F]) were the most difficult ones for trawling. Numerous tows failed to produce valid results. In these strata, the highest proportion of animals was of benthopelagic crustaceans, usually making diel migrations, even though they were normally present on the seafloor. Benthic cnidarians were dominated by octocorals of the Order PENNATULACEA (Verrill, 1865), with a wide bathymetric distribution, adapted to live on soft substrates.

### **3.3. Sediments**

Sediment data obtained during Patagonia 1207, Patagonia 0108, and Atlantis 2008 cruises showed that fine sands were generally predominant throughout the study area, with low contents of organic matter and sediment sorting varying from poor to moderately good. In more detail, the bathymetric sedimentary classification would be as follows:

Depths <200 m: fine sand (mean diameter=210 µm) with low organic matter content (mean value=1.14 %), moderately sorted.

Depths from 201 to 400 m: fine sand (mean diameter ranging from 150 to 189 µm) with low organic matter content (mean value=1.06%), moderately well sorted.

Depths from 401 to 700 m: very fine sand (mean diameter from 110 to 120 µm) recording the highest organic matter content (mean value ranging from 2.23% to 2.35%) and also the highest percentage (up to 44.50%) of silt and clay (<62 µm). Sorting was poor to moderate.

Depths from 701 to 1500 m: fine sand sediments similar to those of the shallowest stratum (mean diameter 160 to 190 µm), with low organic matter contents (mean value ranging from 1.43% to 1.68%). Moderately sorted.

Depths >1501 m: the deepest stratum, located in the bottom of submarine channels and canyons, was characterised by the presence of heterogeneous sediments mainly composed of fine sand (mean diameter=200 µm), with low organic content (mean value=1.68%) and poor sorting. This stratum showed the highest percentage (up to 39.5%) of coarse particles (>500 µm).

### **3.4. Fishery footprint**

Bathymetric strata differences clearly arise by comparing the composition of the sampled

Strata 1 and 8 (<200 m) showed a low catch of benthos (17,209 and 41,202 g. respectively), both in number and diversity. We observed a strong dominance of pectinid molluscs of the Genus *Zygochlamys* (Ihering, 1907) (60.39% of the biomass [Figure 3B]), mainly *Z. patagonica* (King & Broderip, 1832), followed by those of the Genus *Chlamys* (Röding, 1798). Vulnerable organisms were practically unrepresented in these shallower strata, probably due to the bottom trawling

Strata 2 and 9 (201-300 m) recorded the lowest catch in terms of biomass (2121 and 1576 g, respectively). In these strata, detritivorous and opportunistic species were predominant, and the presence of vulnerable organisms was negligible again. Compared to strata 1 and 8, we observed an increase of the benthic cnidarians' biomass values, dominated by gorgonians from Family PRIMNOIDAE (Milne Edwards, 1857) (Octocorallia; Gorgonacea) (80.32% in biomass,

Strata 3 and 10 (301-400 m) were hardly sampled due to the reduced number of valid hauls (3 in stratum 3 and 2 in stratum 10). The low benthic biomass and the negligible presence of vulnerable organisms (Figure 3D) could be attributed to bottom fishing activities, as above‐

Strata 4, 11, 5, 6, and 12 of intermediate depths (401-1000 m, [Figure 3E]) recorded high biomass and numbers of octocorals, sponges, colonial scleractinians (*Bathelia candida*), and large hydrocorals. Octocorals included colonies of various genera belonging to families PRIMNOI‐ DAE and ISIDIDAE. As aforementioned, the increase and proliferation of these species create complex 3D structures providing the ideal habitat for a wide range of organisms. In those strata, the large amount of filter feeders and suspensivore sessile organisms is an indication of the presence of unaltered, complex and structured ecosystems. In the future, ROV and other

Strata 7 and 13 (1001-1500 m, [Figure 3F]) were the most difficult ones for trawling. Numerous tows failed to produce valid results. In these strata, the highest proportion of animals was of benthopelagic crustaceans, usually making diel migrations, even though they were normally present on the seafloor. Benthic cnidarians were dominated by octocorals of the Order PENNATULACEA (Verrill, 1865), with a wide bathymetric distribution, adapted to live on

Sediment data obtained during Patagonia 1207, Patagonia 0108, and Atlantis 2008 cruises showed that fine sands were generally predominant throughout the study area, with low contents of organic matter and sediment sorting varying from poor to moderately good. In

Depths <200 m: fine sand (mean diameter=210 µm) with low organic matter content (mean

Depths from 201 to 400 m: fine sand (mean diameter ranging from 150 to 189 µm) with low

more detail, the bathymetric sedimentary classification would be as follows:

organic matter content (mean value=1.06%), moderately well sorted.

activities for years by bottom trawlers from international fleets.

submersible camera systems could confirm these assumptions.

benthic megafauna (Figures 3B-F):

282 12 Biodiversity in Ecosystems - Linking Structure and Function

[Figure 3C]).

soft substrates.

**3.3. Sediments**

value=1.14 %), moderately sorted.

mentioned for strata 1 and 8.

The statistical analysis of the bottom trawl footprint plot generated with the georeferenced fishery data obtained by the IEO scientific observers (between 1989 and 2007, 9013 fishing operations) showed that most of the commercial hauls of the Spanish fishing fleet in the study area (99.85%) took place at depths below 300 m (Figure 4).

**Figure 4.** Location of commercial hauls and fishery footprint (5'×10') of the Spanish bottom trawl fleet on the HS of the SW Atlantic (1989-2007).

## **4. Multivariate analysis**

### **4.1. Model selection**

Predictive accuracy of the models was evaluated through multiple cross-validation proce‐ dures, splitting the original data three times into two random subsets for calibration (80% data) and evaluation (20% data). The mean area under the receiver operating characteristic (ROC) curve (AUC) obtained from the three repetitions served to assess the predictive performance index of the model. AUC ranks from 0.5 to 1, null accuracy or perfect accuracy of the model, respectively [83]. Table 5 shows the best predictive performance score of the RF model, which was subsequently chosen for vulnerable species modelling.


**Table 5.** Validation of the predictive performance of the four candidate presence/absence models tested (RF: Random Forest; GBM: Generalized Boosting Model; MARS: Multivariate Adaptative Regression Splines; GAM: Generalized Additive Model).

### **4.2. Variable influence**

The Mean Decrease Gini method, implemented in the BIOMOD package, was used to measure the importance of the dependent variable. This exploration tool shows graphically the total decrease in node impurities from splitting on the variable, averaged over all trees. Thus, for classification, the node impurity is measured by the Gini index [72]. The higher the value in the X axis, the higher importance the indicated variable will have on the classification of the dependent variable. Figure 5 show that the topographic position is the main variable affecting the distribution of vulnerable organisms in the HS of the PS, followed by the slope and the sea bottom temperature. Comparatively, the sea floor granulometry has a negligible effect on the distribution of the vulnerable organisms.

In addition to this, it is possible to visualize how each environmental variable, independently from any other, influences the response variable using partial dependence plots [73], which graphically represents the relationships between each predictor variable and the predicted occurrence probabilities of the vulnerable organisms obtained from the RF model. Figure 6 show that bathymetry has a positive effect between 500 and 1000 m depth. Regarding the topographic position, the highest interactions with the presence of vulnerable organisms were observed in canyons (6), followed by abyssal flats (5) and the slope (3 and 4). On the shelf, only outcrop areas (2) were positively correlated with the dependent variable.

A First Approach to Assess the Impact of Bottom Trawling Over Vulnerable Marine Ecosystems on the High Seas… 15 http://dx.doi.org/ 10.5772/59268 285

**4. Multivariate analysis**

284 14 Biodiversity in Ecosystems - Linking Structure and Function

Predictive accuracy of the models was evaluated through multiple cross-validation proce‐ dures, splitting the original data three times into two random subsets for calibration (80% data) and evaluation (20% data). The mean area under the receiver operating characteristic (ROC) curve (AUC) obtained from the three repetitions served to assess the predictive performance index of the model. AUC ranks from 0.5 to 1, null accuracy or perfect accuracy of the model, respectively [83]. Table 5 shows the best predictive performance score of the RF model, which

**Model Mean cross validation score**

**Table 5.** Validation of the predictive performance of the four candidate presence/absence models tested (RF: Random Forest; GBM: Generalized Boosting Model; MARS: Multivariate Adaptative Regression Splines; GAM: Generalized

The Mean Decrease Gini method, implemented in the BIOMOD package, was used to measure the importance of the dependent variable. This exploration tool shows graphically the total decrease in node impurities from splitting on the variable, averaged over all trees. Thus, for classification, the node impurity is measured by the Gini index [72]. The higher the value in the X axis, the higher importance the indicated variable will have on the classification of the dependent variable. Figure 5 show that the topographic position is the main variable affecting the distribution of vulnerable organisms in the HS of the PS, followed by the slope and the sea bottom temperature. Comparatively, the sea floor granulometry has a negligible effect on the

In addition to this, it is possible to visualize how each environmental variable, independently from any other, influences the response variable using partial dependence plots [73], which graphically represents the relationships between each predictor variable and the predicted occurrence probabilities of the vulnerable organisms obtained from the RF model. Figure 6 show that bathymetry has a positive effect between 500 and 1000 m depth. Regarding the topographic position, the highest interactions with the presence of vulnerable organisms were observed in canyons (6), followed by abyssal flats (5) and the slope (3 and 4). On the shelf, only

outcrop areas (2) were positively correlated with the dependent variable.

RF 0.876 GBM 0.825 MARS 0.804 GAM 0.778

was subsequently chosen for vulnerable species modelling.

**4.1. Model selection**

Additive Model).

**4.2. Variable influence**

distribution of the vulnerable organisms.

**Figure 5.** Mean Decrease Gini for each explanatory variable in the RF model. Higher values in the X axis indicate high‐ er influence of the environmental variable on the occurrence of benthic vulnerable organisms.

**Figure 6.** Partial dependence plots showing quantitative influence from each environmental variable on the occurrence of benthic vulnerable organisms predicted probability.

### **4.3. HSI mapping**

Table 6 shows the presence data of benthic vulnerable organisms from the 169 sampled locations. Predicted values were plotted to produce a habitat suitability map showing survey sampling stations with presence/absence data and the vulnerable organisms' probability of occurrence (Figure 7).


**Table 6.** Summary of the presence sampling data of vulnerable organisms (VO).

**Figure 7.** HSI map of benthic vulnerable organisms. Higher probability of occurrence is shown in darker tones. Survey sampling stations are overlapped, showing presence (black dot) or absence (circle) of such organisms.

## **5. Conclusions**

**4.3. HSI mapping**

286 16 Biodiversity in Ecosystems - Linking Structure and Function

occurrence (Figure 7).

Table 6 shows the presence data of benthic vulnerable organisms from the 169 sampled locations. Predicted values were plotted to produce a habitat suitability map showing survey sampling stations with presence/absence data and the vulnerable organisms' probability of

**Figure 7.** HSI map of benthic vulnerable organisms. Higher probability of occurrence is shown in darker tones. Survey

sampling stations are overlapped, showing presence (black dot) or absence (circle) of such organisms.

**Organism Presence** Alcyonacea 15 *Bathelia candida* 23 Demospongiae 22 Gorgonacea 24 Hexactinellidae 14 Hydrozoa 41 Pennatulacea 7 Rhodalidae 9 Stylasteridae 25 **Total VO 76**

**Table 6.** Summary of the presence sampling data of vulnerable organisms (VO).

Multibeam acoustic data showed that the upper slope and uppermost middle slope were scarred by iceberg plough marks. The middle slope surveyed was entrenched by seven submarine canyons [78]. Pockmarks and other seismic and morphologic evidence of gas/fluids seepage were pervasive throughout the entire survey area and more intense in the southern middle part [80]. Water coral communities associated with those pockmarks could be quite extensive in the study area.

The highest benthic biodiversity was found between 800 and 1500 m depth. Biodiversity was higher along the continental margin (per an equal number of individuals, and in terms of abundance) than biodiversity found along the continental shelf. Our results have confirmed the existence of close ecological relationships between Patagonian deep-sea fauna and Antarctic fauna of shallow waters. Benthic megafauna collected included invertebrates, chordates, and hemichordates. There was a clear dominance in biomass and diversity of the Phyla Porifera and Cnidaria. Most species of these groups are considered as vulnerable according to UN and OSPAR criteria: sponges, octocorals, colony scleractinians, anthipatari‐ ans, calcified hydrozoans (Family STYLASTERIDAE), and erect bryozoans (Table 7).

Shallow waters (<400 m) are the strata having sustained most of the fishing pressure for almost 50 years. Below 400 m we recorded the lowest biomass, abundance and diversity values, most likely due to this fishing pressure. In these strata we noted the presence of sparse organisms with erected growth, a high dominance of pectinid mollusks (*Zygochlamys patagonica*), and minor presence of species considered as indicators of VMEs.

Intermediate depths (401-1000 m) showed an important increase in number and biomass of vulnerable organisms, with outstanding numbers, densities and biomasses of octocorals, sponges, colony scleractinians (*Bathelia candida*) and big hydrocorals (*Errina* spp., *Cheiloporidion pulvinatum* [Cairns, 1983], *Sporadopora* sp., and *Stylaster densicaulis* [Moseley, 1879]). Also remarkable was the presence of sponges of the Family CLADORHIZIDAE (Dendy, 1922), a group of a great zoological importance because they are carnivorous and have developed a trophic adaptation to live in the ocean's depths.

In deeper strata (1001-1500 m) we found more anomuran crustaceans of the Family LITHO‐ DIDAE (Samouelle, 1819) (mainly *Paralomis formosa* [Henderson, 1888]). Amongst benthic cnidarians, the pennatulid octocorals (Order PENNATULACEA) were the most abundant.

The model accuracy is acceptable (0.876). Although the modelling' accuracy values were higher when considering each organism, this was an expected fact due to the different environmental preferences of the studied organisms. However, HSI mapping is a useful conservation management tool enabling an initial observation of how environmental condi‐ tions control the spatial distribution of vulnerable organisms in the study area. The research will proceed further when data from all 13 survey cruises undertaken in the area are analysed.

The main environmental conditions affecting presence of vulnerable organisms seems to be connected to the topographic position, slope and bathymetry. Sea bed granulometry appeared


**Table 7.** Cold-water corals and deep-water sponges concentrations: list of most common species collected in the campaigns of the Atlantis Project in 2007 and 2008.

to have a negligible effect on the presence of vulnerable organisms, contradicting published research results on this subject in other geographical areas, where substrate characteristics determine to a large extent the presence or absence of a particular benthic species [84-86].

**Porifera Grant, 1836** *Chondrocladia* sp. **Class Hexactinellida Schmidt, 1870** *Euchelipluma* sp.

288 18 Biodiversity in Ecosystems - Linking Structure and Function

*Cynachyra* sp *Latrunculia* sp. *Geodia* sp. Axinellidae indet.

Lithistidindet. *Errina* sp.

*Raspailia* sp. *Alcyonium* sp. *Inflatella* sp. *Anthomastus* sp. *Pyloderma latrunculioides* Ridley & Dendy, 1886 *Paragorgia* sp. *Desmacidon*, sp. *Primnoella* sp. *Hymedesmia* (*Hymedesmia*) sp. Isididae indet. *Hymedesmia* (*Stylopus*) sp. *Anthoptilum* sp. *Phorbas* sp. *Halipteris* sp. *Myxilla* (*Myxilla*) *mollis* Ridley & Dendy, 1886 *Epizoanthus* sp.

*Iophon* sp. *Stylaster* cf. *densicaulis*

*Clathria* sp. **Class Anthozoa Ehrenberg, 1831**

*Myxilla* (*Burtonanchora*) *lissostyla* Burton, 1938 *Actinostola crassicornis* Hertwig, 1882

**Table 7.** Cold-water corals and deep-water sponges concentrations: list of most common species collected in the

*Tedania* (*Tedaniopsis*) *charcoti* Topsent, 1907 *Bathelia candida* Moseley, 1881

*Tedania* (*Tedaniopsis*) *oxeata* Topsent, 1916 *Caryophyllia* sp. *Tedania* (*Tedaniopsis*) *massa* Ridley & Dendy, 1886 *Desmophyllum* sp. *Tedania* (*Trachytedania*) sp. *Flabellum* sp.

campaigns of the Atlantis Project in 2007 and 2008.

*Asbestopluma* sp.

*Polymastia* sp. *Haliclona* (*Haliclona*) sp. *Radiella* sp. *Haliclona* (*Gellius*) sp. *Tentorium* sp. Dictyoceratida indet. *Stylocordyla* cf. *stipitata* **Cnidaria Hatschek, 1888** *Timea* sp. **Class Hydrozoa Owen, 1843**

*Rossella antarctica* Carter, 1872 *Mycale* (*Oxymycale*) *acerata* Kirkpatrick, 1907 **Class Demospongiae Sollas, 1885** *Mycale* (*Carmia*) *gaussiana* Hentschel, 1914 *Tetilla leptoderma* Sollas, 1886 *Isodictya kerguelenensis* Ridley & Dendy, 1886

Our study only calculated the general trends of the granulometrical parameters, while bathymetry, slope and topographic position were variables derived from high resolution data, strongly correlated with the response variable. Therefore, local conditions are the main factors ruling the potentiality of a habitat to host benthic vulnerable organisms in the HS of the PS.

The use of the Random Forest model offers both higher classification accuracy and determi‐ nation of variable importance, and more stability where small perturbations of the data exist [76]. RF is a predictive classification and algorithm that does not make any distributional assumptions about the predictor or the response variables. It also handles situations in which the number of predictor variables exceeds the number of observations, offering a powerful non parametrical alternative for ecological modelling [64].

The vulnerable species groups, communities and habitats described here are mainly distrib‐ uted beyond the 500 m depth contour. The presence of organisms considered as vulnerable is almost negligible in the fishing area. This fact is almost certainly due to bottom trawling operations of international fleets taking place in the study area for nearly 50 years. Also, the fishing grounds are far away from the geographical location of the main geomorphological features such as canyons, trenches, gas and fluid seepages observed in the middle slope, and identified as potential sites for VMEs.

The fishery footprint plot shows that the historical activity of the Spanish bottom trawler fleet has been located in the shallowest depth strata, at depths not generally exceeding 300 m. On this basis we think that the adverse impacts of current bottom fishing activities on VMEs are negligible or small. However, the displacement of the fishing fleet to target deep sea species at greater depths (were the existence of VMEs has been observed) could have a negative impact on those ecosystems. With this in mind and following the FAO deep-water guidelines, the potential threat of such a fishing strategy should be assessed.

Apart from Spanish fishing fleet, other bottom trawling fleets from different nations (former Soviet Union, Poland, GDR, Bulgaria, etc) have been operating intensively in the SW Atlantic (including our study area) from mid 60's until mid 80's, both over the continental shelf and slope [87-92]. Even if no data were made available to us for assessing the eventual negative impact of these fleets on VMEs, some experiences in other geographical areas such as the North Atlantic, Southwest and East Pacific, seamounts off Tasmania, and waters off New Zealand [31,45,92], have shown that high fishing pressure exerted by a large number of bottom trawlers over a long period of time could relevantly affect these VMEs. We therefore think that probably the almost 50 years of intensive bottom trawling in this SW Atlantic area by the abovemen‐ tioned fleets could have contributed to the low presence of VMEs in the study area at depths lower than 500 m.

## **Acknowledgements**

We wish to thank the crew of the R/V "Miguel Oliver" (owned by the Spanish General Secretariat for Fisheries [SGP]) and her captain, for the professionalism and the courtesy extended towards us during the research cruises. We are also very grateful to all those involved in the five research campaigns, namely the scientific and technical personnel who made this work possible, known as the Atlantis Group: J. M. Cabanas, J. Gago, B. Almón, E. Elvira, P. Jiménez, A. Fontán, C. Alcalá, and V. López, among others. Our thanks also to the Spanish Secretaría General del Mar (SGM, General Secretariat for the Sea), owner of the research ship, for giving us the opportunity to conduct this research.

## **Author details**

J. Portela1\*, J. Cristobo2 , P. Ríos2 , J. Acosta3 , S. Parra4 , J.L. del Río1 , E. Tel3 , V. Polonio2 , A. Muñoz5 , T. Patrocinio1 , R. Vilela1 , M. Barba1 and P. Marín5

\*Address all correspondence to: julio.portela@vi.ieo.es

1 Spanish Institute of Oceanography (IEO), Vigo Oceanographic Center. Fisheries Dept. Vigo, Spain

2 Spanish Institute of Oceanography, Gijón Oceanographic Center. Benthos Dept. Gijón, Spain

3 Spanish Institute of Oceanography, Headquarters Madrid. Geology Dept. Madrid, Spain

4 Spanish Institute of Oceanography, A Coruña Oceanographic Center. Epibenthos Dept. A Coruña, Spain

5 Multidisciplinary Mapping Group (TRAGSATEC-SGP), Spanish General Secretariat for Fisheries (SGP). Cartography Dept. Madrid, Spain

### **References**


[3] Portela, J.M., Pierce, G.J., del Río, J.L., Sacau, M., Patrocinio, T., Vilela, R. Preliminary description of the overlap between squid fisheries and VMEs on the high seas of the PatagonianShelf. Fisheries Research, 106 (2010) 229–238.

**Acknowledgements**

290 20 Biodiversity in Ecosystems - Linking Structure and Function

**Author details**

T. Patrocinio1

Vigo, Spain

Coruña, Spain

**References**

Spain

J. Portela1\*, J. Cristobo2

for giving us the opportunity to conduct this research.

, P. Ríos2

\*Address all correspondence to: julio.portela@vi.ieo.es

Fisheries (SGP). Cartography Dept. Madrid, Spain

Paper 457. Rome, FAO. 2005. 235 pp.

No. 522. Rome, FAO. 2008. 145p.

, M. Barba1

, R. Vilela1

, J. Acosta3

We wish to thank the crew of the R/V "Miguel Oliver" (owned by the Spanish General Secretariat for Fisheries [SGP]) and her captain, for the professionalism and the courtesy extended towards us during the research cruises. We are also very grateful to all those involved in the five research campaigns, namely the scientific and technical personnel who made this work possible, known as the Atlantis Group: J. M. Cabanas, J. Gago, B. Almón, E. Elvira, P. Jiménez, A. Fontán, C. Alcalá, and V. López, among others. Our thanks also to the Spanish Secretaría General del Mar (SGM, General Secretariat for the Sea), owner of the research ship,

, S. Parra4

1 Spanish Institute of Oceanography (IEO), Vigo Oceanographic Center. Fisheries Dept.

2 Spanish Institute of Oceanography, Gijón Oceanographic Center. Benthos Dept. Gijón,

3 Spanish Institute of Oceanography, Headquarters Madrid. Geology Dept. Madrid, Spain

4 Spanish Institute of Oceanography, A Coruña Oceanographic Center. Epibenthos Dept. A

5 Multidisciplinary Mapping Group (TRAGSATEC-SGP), Spanish General Secretariat for

[1] FAO. Review of the state of world marine fishery resources. FAO Fisheries Technical

[2] Bensch, A., Gianni, M., Gréboval, D., Sanders, J.S., Hjort, A. Worldwide review of bottom fisheries in the high seas. FAO Fisheries and Aquaculture Technical Paper.

and P. Marín5

, J.L. del Río1

, E. Tel3

, V. Polonio2

, A. Muñoz5

,


ments. 2003. Available from: http://daccess-dds-ny.un.org/doc/UNDOC/GEN/ N03/453/75/PDF/N0345375.pdf? OpenElement, 12pp.

[14] Pierce, J.G., Portela, J. Fisheries Production and Market Demand. In: J. Iglesias et al. (eds.) Cephalopod Culture, DOI 10.1007/978-94-017-8648-5\_3, © Springer Science

[15] Maguire, J.J., Sissenwine, M., Csirke, J., Grainger, R., Garcia, S. The state of world highly migratory, straddling and other high seas fishery resources and associated

[16] Kaiser M.J. Significance of Bottom-Fishing Disturbance. Conservation Biology, pages

[17] Kaiser, M.J., Edwards, D.B., Armstrong, P.J., Radford, K., Lough, N.E.L., Flatt, R.P., Jones H.L. Changes in megafaunal benthic communities in different habitats after trawling disturbance. ICES (International Council for the Exploration of the Sea)

[18] Collie, J.S., Escanero, G.A., Valentine, P.C. Photographic evaluation of the impacts of bottom fishing on benthic epifauna. ICES Journal of Marine Science 2000, 57: 987–

[19] Collie, J.S., Hall, S.J., Kaiser, M.J., Poiner I.R. A quantitative analysis of fishing im‐ pacts on shelf-sea benthos. Journal of animal ecology (2000) 69, 785-798. 2000b. [20] Coggan, R., Populus, J., White, J., Sheehan, K., Fitzpatrick, F. & Piel, S. (Eds.). Review of Standards and Protocols for Seabed Habitat Mapping, 2007. MESH project docu‐

[21] Durán Muñoz, P., Murillo, F.J., Serrano, A., Sayago-Gil, M., Parra, S., Díaz Del Río, V., Sacau, M., Patrocinio, T. & Cristobo, J. A Case Study of available methodology for the identification of Vulnerable Ecosystems/Habitats in bottom deep-sea fisheries: Possibilities to apply this method in the NAFO Regulatory Area in order to select

Marine Protected Areas. NAFO SCR Doc. 08/06 Serial No. N5491, 2008. 20 pp.

ICES Journal of Marine Science. 2009, 66: 2013–2025.

rine Science. 2010, 68 (2): 319-332.

[22] Durán Muñoz, Sayago-Gil, M., Cristobo, J., Parra, S., Serrano, A., Díaz Del Río, V., Patrocinio, T., Murillo, F. J., Palomino, D., & Fernández-Salas, L.M. Seabed mapping for selecting cold-water coral protection areas on Hatton Bank, Northeast Atlantic. –

[23] Murillo, F. J., Durán Muñoz, P., Altuna, A. & Serrano, A. Distribution of deep-water corals of the Flemish Cap, Flemish Pass, and the Grand Banks of Newfoundland (Northwest Atlantic Ocean): interaction with fishing activities. ICES Journal of Ma‐

[24] UNGA. Resolution 58/14. Sustainable fisheries, including through the 1995 Agree‐ ment for the Implementation of the Provisions of the United Nations Convention on the Law of the Sea of 10 December 1982 relating to the Conservation and Manage‐ ment of Straddling Fish Stocks and Highly.Migratory Fish Stocks, and related instru‐

species. FAO Fisheries Technical Paper 495. Rome, FAO, 2006. 84 pp.

+Business Media Dordrecht 2014.

292 22 Biodiversity in Ecosystems - Linking Structure and Function

1230–1235. Volume 12, No. 6, December 1998.

Journal of Marine Science, 1998, 55: 353–362.

1001. 2000a.

ment.


ment Australia and the Fisheries Research and Development Corporation, 1998. Rep. FRDC Project 95/058, CSIRO, Hobart, Tasmania, Australia, 104pp.

[44] Koslow, J.A., Gowlett-Holmes, K., Lowry, J.K., O'Hara, T., Poore, G.C.B., Williams, A. Seamount benthic macrofauna off southern Tasmania: community structure and impacts of trawling. Marine Ecology Progress Series 2001, 213: 111–125.

[33] Hiddink, J. G., Jennings, S., Kaiser, M.J. Assessing and predicting the relative ecologi‐ cal impacts of disturbance on habitats with different sensitivities. Journal of Applied

[34] Nellemann, C., Hain, S., Alder, J. (Eds.). February 2008. In Dead Water – Merging of climate change with pollution, over-harvest, and infestations in the world's fishing grounds. United Nations Environment Programme, GRID-Arendal, Norway,

[35] Weaver, P.P.E., Benn, A., Arana, P.M., Ardron, J.A., Bailey, D.M., Baker, K., Billett, D.S.M., Clark, M.R., Davies, A.J., Durán Muñoz, P., Fuller, S.D., Gianni, M., Grehan, A.J., Guinotte, J., Kenny, A., Koslow, J.A., Morato, T., Penney, A.J., Perez, J.A.A., Priede, I.G., Rogers, A.D., Santos, R.S., Watling, L. The impact of deep-sea fisheries and implementation of the UNGA Resolutions 61/105 and 64/72. Report of an inter‐ national scientific workshop, National Oceanography Centre, Southampton, (2011)

[36] Hall-Spencer, J., Allain, V., Fosså, J.H. Trawling damage to Northeast Atlantic an‐ cient coral reefs. Proceedings of the Royal Society of London, 2002, Series B: Biologi‐

[37] Fosså, J., Mortensen, P., Furevik, D. The deep-water coral *Lophelia pertusa* in Norwe‐ gian waters: distribution and fishery impacts. Hydrobiologia (2002) 471: 1–12. [38] Wheeler, A.J., Bett, B.J., Billett, D.S., Masson, D.G., Mayor, D. The impact of demersal trawling on northeast Atlantic coral habitats: The case of the Darwin mounds, United Kingdom. In: Barnes, P.W., Thomas, J.P. (Eds.) Benthic habitats and the Effects of fishing, American Fisheries Society Symposium 41, 2005, Bethesda, Maryland, USA,

[39] Mortensen, P.B., Buhl-Mortensen, L. Gordon Jr, D.C. Fader, G.B.J., McKeown, D.L., Fenton, D.G. Effects of Fisheries on Deep-water Gorgonian Corals in the Northeast Channel, Nova Scotia (Canada). American Fisheries Society Symposium 2005, 41:

[40] Edinger, E., Baker, K., Devillers, R., Wareham, V.E. Coldwater Corals off Newfound‐ land and Labrador: Distribution and Fisheries Impacts. WWF-Canada, Toronto, Can‐

[41] Stone, R.P. Coral habitat in the Aleutian Islands of Alaska: depth distribution, finescale species associations, and fisheries interactions. Coral Reefs, 2006, 25: 229–238. [42] Krieger, K.J. Coral (Primnoa) impacted by fishing gear in the Gulf of Alaska. In: Wil‐ lison. J.H.M., Gass, S.E., Kenchington, E.l.R., et al. (Eds.) Proceedings of the first In‐ ternational Symposium on Deep-Sea Corals, Ecology Action Centre and Nova Scotia

[43] Koslow, J.A., Gowlett-Holmes, K. The seamount fauna off southern Tasmania: benth‐ ic communities, their conservation and impacts of trawling: Final Report to Environ‐

museum, Halifax, Nova Scotia, Canada, 2001, pp. 106–116.

45 pp. Available from http://hdl.handle.net/10013/epic.37995.

Ecology (2007) 44 (2): 405-413.

294 24 Biodiversity in Ecosystems - Linking Structure and Function

cal Sciences 269(1490): 507–511.

www.grida.no.

pp. 807–817.

369-382.

ada, 2007, 41pp.


[69] R Development Core Team. R: A language and environment for statistical comput‐ ing. R Foundation for Statistical Computing 2011, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/.

[56] WGEAFM. Report of the NAFO SC Working Group on Ecosystem Approach to Fish‐ eries Management (WGEAFM). NAFO Scientific Council Meeting June 2008. Serial

[57] WGEAFM. Report of the NAFO SC Working Group on Ecosystem Approach to Fish‐ eries Management (WGEAFM). Response to Fisheries Commission Request 9.a. NA‐ FO Scientific Council Meeting October 2008, Serial Number 5592, NAFO SCS Doc.

[58] Rogers, A.D., Gianni, M. The Implementation of UNGA Resolutions 61/105 and 64/72 in the Management of Deep-Sea Fisheries on the High Seas. Report prepared for the Deep-Sea Conservation Coalition. International Programme on the State of the

[59] ICES. Report of the Working Group on Marine Habitat Mapping (WGMHM), 21-24 April 2009, The National Institute of Aquatic Resources, Charlottenlund Castle, Co‐

[60] NAFO. Report of the NAFO SC Working Group on Ecosystem Approach to Fisheries Management (WGEAFM). Response to Fisheries Commission Request 9.a. NAFO

[61] Cutler, D.R., Edwards, T.C., Beard, K.H., Cutler, A., Hess, K.T., Gibson, J., Lawler, J.J. Random forests for classification in ecology. Ecology 2007, 88, 2783–2792.

[62] Crisci, C., Ghattas,B., Perera, G. A review of supervised machine learning algorithms and their applications to ecological data. Ecological Modelling 2012, 240, 113-122. [63] Prasad, A.M., Iverson, L.R., Liaw, A. Newer classification and regression tree techni‐ ques: bagging and random forest for ecological prediction. Ecosystems 2006. 9:

[64] Hirzel, A.H., Hausser, J., Chessel, D., Perrin, N. Ecological-niche factor analysis: How to compute habitat-suitability maps without absence data? Ecology, 83(7), 2002, pp.

[65] Kuhnert, P.M., Martin, T.G., Mengersen, K., Possingham, H.P. Assessing the impacts of grazing levels on bird density in woodland habitat: a Bayesian approach using ex‐

[66] Zuur, A.F., Ieno, E.N., Meesters, E.H.W.G. A Beginner's Guide to R. Springer, 2009,

[67] Breiman, L. Random Forests. (S. R. E, Ed.). Machine Learning 2001. 45(1), 5-32.

[68] Kohavi R., Provost, F. Glossary of terms. Machine Learning 1998. 30, (2-3): 271-274.

Number 5511, NAFO SCS Doc. 08/10, 70pp. 2008a

Ocean, London, United Kingdom, 2010, 97pp.

penhagen, Denmark. ICES CM 2009/MHC: 07. 78 pp.

SCS Document, 2008, 08/24, Serial No. N5592. 19 pp.

pert opinion. Environmetrics, 2005, 16: 717–747.

08/24, 19pp. 2008b

296 26 Biodiversity in Ecosystems - Linking Structure and Function

181-199.

2027-2036.

New York.

Springer.


## **Agrorural Ecosystem Effects on the Macroinvertebrate Assemblages of a Tropical River**

Bert Kohlmann, Alejandra Arroyo, Monika Springer and Danny Vásquez

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/59073

## **1. Introduction**

Costa Rica is an ideal reference point for global tropical ecology. It has an abundance of tropical forests, wetlands, rivers, estuaries, and active volcanoes. It supports one of the highest known species density (number of species per unit area) [1, 2] on the planet and possesses about 4 % of the world´s total species diversity [3]. Because of its tropical setting, it also serves as an important location for agricultural production, including cultivars such as coffee, bananas, palm hearts, and pineapples. The country has also attracted more ecotourists and adventure travelers per square kilometer than any other country in the world [4].

The agrorural frontier on the Caribbean side of Costa Rica started to spread during the 1970s, especially in its northeastern area. Migrations of land-poor people from the Pacific and mountain areas of the country started to colonize the land that the government had made available [5, 6]. These waves of immigrants tended to establish themselves along river systems. In this way, towns, small to medium-scale family farming, ranching, and plantation agriculture began to base themselves along the main river systems. It was during this time that the human settlements originated along the Dos Novillos River [7].

Residual waters produced by all of the aforementioned human activities are at present discharged into the river systems in the Costa Rican Caribbean area. Households not situated in the neighborhood of rivers will use septic tanks; homesteads situated along riverbanks will discharge their effluents directly into the rivers. Other activities like the production of residual waters from dairy farms, pigsties, banana packing plants, plantations´ excess fertilization, etc. will drain eventually into a river. The Dos Novillos River is no exception. Water sewage systems in this part of Costa Rica are almost non-existent.

© 2015 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2015 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Aquatic biomonitoring in Latin America started at the end of the last century, commencing in Colombia [8-10] and then spreading to other Latin American countries. Revisions of the use of aquatic biomonitoring indices in Latin America are given in de la Lanza Espino *et al*. [8], Prat *et al*. [9], and Springer [10]. In the case of Costa Rica, several studies based on the impor‐ tance of macroinvertebrates for biomonitoring water quality and community structure, and function in banana and conventional and organic rice systems [14-22], as well as macroinver‐ tebrate field-guides [23, 24] have been published. Some studies of macroinvertebrate assem‐ blage structures also have been reported for rivers on the southern Caribbean coast of Costa Rica [25-34].

Although ecosystem studies have been used in Costa Rica for evaluating possible impacts caused by crop activities on river water quality at specific points, almost nothing is known about how a mix of other human activities can impact macroinvertebrate biodiversity and ecosystem structure along the length of a river. The aim of this study was to use macroinver‐ tebrate biodiversity under the influence of different human activities along the length of a river in order to describe their impact on community structure and function in tropical agrorural environments.

## **2. Materials and methods**

### **2.1. Study area**

This study was conducted at the Dos Novillos River (Figure 1, Table 1) in the province of Limón, Costa Rica. Samples were collected two times a month from January 2005 to March 2005 and monthly from April 2005 to January 2006. The Dos Novillos River drains from the Central Cordillera at an elevation of approximately 2380 masl towards the Caribbean lowlands of the province of Limón. This river is part of the 2950 km2 Parismina River watershed in a premontane wet forest and tropical moist forest region [35]. The underlying geology is represented by quaternary sedimentary and volcanic rocks under the influence of nearby volcanoes, with a flat to undulating topography and poorly drained alluvial soils susceptible to flooding [36]. Banana plantations have been developed on the lower reaches of this water‐ shed. The study area is characterized by a humid tropical climate with a mean temperature of 25.8 °C, an average annual relative humidity of 87 %, and an annual precipitation average of 3460 mm ± 750 mm without a pronounced dry season. The sampling area runs in a straight line through the towns and localities of La Argentina, Pocora, and EARTH University in the Province of Limón. Each fore mentioned landmark is separated by approximately 4 km. The total sampling area runs along a length of 13.6 km, across an area with a mixture of a premon‐ tane wet forest, pastureland, small town, riparian tropical moist forest, and banana agricultural areas.

Six sampling sites were located along the Dos Novillos River (Figs. 1-7; Table 1) where macroinvertebrates were sampled. The first sampling site (Figure 2) (site 1, "Don Eladio") served as a reference site, being part of the rhithral region, located upstream of the first anthropogenic disturbance (pastureland). This site is surrounded by tropical rain forest;

Aquatic biomonitoring in Latin America started at the end of the last century, commencing in Colombia [8-10] and then spreading to other Latin American countries. Revisions of the use of aquatic biomonitoring indices in Latin America are given in de la Lanza Espino *et al*. [8], Prat *et al*. [9], and Springer [10]. In the case of Costa Rica, several studies based on the impor‐ tance of macroinvertebrates for biomonitoring water quality and community structure, and function in banana and conventional and organic rice systems [14-22], as well as macroinver‐ tebrate field-guides [23, 24] have been published. Some studies of macroinvertebrate assem‐ blage structures also have been reported for rivers on the southern Caribbean coast of Costa

Although ecosystem studies have been used in Costa Rica for evaluating possible impacts caused by crop activities on river water quality at specific points, almost nothing is known about how a mix of other human activities can impact macroinvertebrate biodiversity and ecosystem structure along the length of a river. The aim of this study was to use macroinver‐ tebrate biodiversity under the influence of different human activities along the length of a river in order to describe their impact on community structure and function in tropical agrorural

This study was conducted at the Dos Novillos River (Figure 1, Table 1) in the province of Limón, Costa Rica. Samples were collected two times a month from January 2005 to March 2005 and monthly from April 2005 to January 2006. The Dos Novillos River drains from the Central Cordillera at an elevation of approximately 2380 masl towards the Caribbean lowlands

premontane wet forest and tropical moist forest region [35]. The underlying geology is represented by quaternary sedimentary and volcanic rocks under the influence of nearby volcanoes, with a flat to undulating topography and poorly drained alluvial soils susceptible to flooding [36]. Banana plantations have been developed on the lower reaches of this water‐ shed. The study area is characterized by a humid tropical climate with a mean temperature of 25.8 °C, an average annual relative humidity of 87 %, and an annual precipitation average of 3460 mm ± 750 mm without a pronounced dry season. The sampling area runs in a straight line through the towns and localities of La Argentina, Pocora, and EARTH University in the Province of Limón. Each fore mentioned landmark is separated by approximately 4 km. The total sampling area runs along a length of 13.6 km, across an area with a mixture of a premon‐ tane wet forest, pastureland, small town, riparian tropical moist forest, and banana agricultural

Six sampling sites were located along the Dos Novillos River (Figs. 1-7; Table 1) where macroinvertebrates were sampled. The first sampling site (Figure 2) (site 1, "Don Eladio") served as a reference site, being part of the rhithral region, located upstream of the first anthropogenic disturbance (pastureland). This site is surrounded by tropical rain forest;

Parismina River watershed in a

of the province of Limón. This river is part of the 2950 km2

Rica [25-34].

environments.

**2.1. Study area**

areas.

**2. Materials and methods**

3002 Biodiversity in Ecosystems - Linking Structure and Function

**Figure 1.** Sampling site location and areas of major potential anthropogenic disturbance at the Dos Novillos River, Guácimo, Limón, Costa Rica [modified from [37].

therefore, natural good water conditions were expected, as well as high taxa richness and an assemblage composition dominated by pollution-sensitive organisms. Site 2, "La Argentina" (Figure 3), was located approximately 5.5 km downstream from site 1. This site was selected

Figure 2. The first site (Don Eladio) was established as a control (reference) site and is located several miles upstream of La Argentina, Pocora, inside natural tropical forest. The site is characterized by a variety of current conditions, with fast flowing riffles, medium laminar flow and pools. The substrate is composed mostly of medium-sized rocks, many of them covered with moss, though large boulders also stand, where large numbers of insects live in the splash area. **Figure 2.** The first site (Don Eladio) was established as a control (reference) site and is located several miles upstream of La Argentina, Pocora, inside natural tropical forest. The site is characterized by a variety of current conditions, with fast flowing riffles, medium laminar flow and pools. The substrate is composed mostly of medium-sized rocks, many of them covered with moss, though large boulders also stand, where large numbers of insects live in the splash area.

to examine the possible extent that small livestock farming might have on river water quality. Site 3, "Chiquitín" (Figure 4), was located approximately 8 km downstream from site 1. High anthropogenic influence was expected at this site because the houses situated at the riverfront discharge their grey and black waters directly into the river. Site 4, "Puente La Hamaca" (Figure 5), was located within the property of EARTH University, approximately 2 km downstream Agrorural Ecosystem Effects on the Macroinvertebrate Assemblages of a Tropical River 5 http://dx.doi.org/ 10.5772/59073 303

Figure 3. Site 2 (La Argentina) is located downstream of La Argentina. It is a place dominated by rocks of all sizes, although big boulders are less numerous than site 1. Current conditions are similar as in Site 1. **Figure 3.** Site 2 (La Argentina) is located downstream of La Argentina. It is a place dominated by rocks of all sizes, although big boulders are less numerous than site 1. Current conditions are similar as in Site 1.

from site 3. As the intervening river length between sites 3 and 4 runs through forest areas, site 4 was selected in order to examine if water quality was improved by a forest filtering processes. Site 5, "Desembocadura" (Figure 6), is within the EARTH University campus and is located approximately 500 m upstream from the confluence of the Dos Novillos and Parismina Rivers. Site 5 was selected to analyze the impact of banana plantations on river water quality. Site 6, "Quebrada Mercedes" (Figure 7), is one tributary of the Dos Novillos River

to examine the possible extent that small livestock farming might have on river water quality. Site 3, "Chiquitín" (Figure 4), was located approximately 8 km downstream from site 1. High anthropogenic influence was expected at this site because the houses situated at the riverfront discharge their grey and black waters directly into the river. Site 4, "Puente La Hamaca" (Figure 5), was located within the property of EARTH University, approximately 2 km downstream

**Figure 2.** The first site (Don Eladio) was established as a control (reference) site and is located several miles upstream of La Argentina, Pocora, inside natural tropical forest. The site is characterized by a variety of current conditions, with fast flowing riffles, medium laminar flow and pools. The substrate is composed mostly of medium-sized rocks, many of them covered with moss, though large boulders also stand, where large numbers of insects live in the splash area.

stand, where large numbers of insects live in the splash area.

3024 Biodiversity in Ecosystems - Linking Structure and Function

Figure 2. The first site (Don Eladio) was established as a control (reference) site and is located several miles upstream of La Argentina, Pocora, inside natural tropical forest. The site is characterized by a variety of current conditions, with fast flowing riffles, medium laminar flow and pools. The substrate is composed mostly of medium-sized rocks, many of them covered with moss, though large boulders also flowing through a forested area within the EARTH University campus, approximately 2 km downstream from site 3. Site 6 was chosen to examine if the intermittent discharge of a small drain of water used to wash bananas in a packing plant had any effect on the stream. Table 1 indicates the exact location, depth, width, and current conditions for each site.

current is more laminar; the substrate is also rocky, but there is no presence of large boulders. **Figure 4.** This site (Chiquitín) is located in the center of the town of Pocora. It is at this point where the channel be‐ comes wider and current is more laminar; the substrate is also rocky, but there is no presence of large boulders.

Figure 4. This site (Chiquitín) is located in the center of the town of Pocora. It is at this point where the channel becomes wider and

flowing through a forested area within the EARTH University campus, approximately 2 km downstream from site 3. Site 6 was chosen to examine if the intermittent discharge of a small drain of water used to wash bananas in a packing plant had any effect on the stream. Table 1

Figure 4. This site (Chiquitín) is located in the center of the town of Pocora. It is at this point where the channel becomes wider and

current is more laminar; the substrate is also rocky, but there is no presence of large boulders. **Figure 4.** This site (Chiquitín) is located in the center of the town of Pocora. It is at this point where the channel be‐ comes wider and current is more laminar; the substrate is also rocky, but there is no presence of large boulders.

indicates the exact location, depth, width, and current conditions for each site.

3046 Biodiversity in Ecosystems - Linking Structure and Function

forest borders the channel at this point. **Figure 5.** Site 4 (La Hamaca) is located within the EARTH University campus, specifically at the suspension bridge. The diversity of current conditions is similar to the other sites, although the rock size is much smaller as at the other three upstream sites. The gallery forest borders the channel at this point.

Figure 5. Site 4 (La Hamaca) is located within the EARTH University campus, specifically at the suspension bridge. The diversity of current conditions is similar to the other sites, although the rock size is much smaller as at the other three upstream sites. The gallery

trees dominate vegetation along the channel. At this site, small airplanes were observed flying over the river while spraying pesticides on the surrounding banana plantations. **Figure 6.** Site 5 (Desembocadura) corresponds to the mouth of the River Dos Novillos with the Parismina. The sub‐ strate consists almost entirely of sand. Big boulders are absent, small rocks are scarce and current flow is weak; tall grasses, banana plants, bamboo, and a few trees dominate vegetation along the channel. At this site, small airplanes were observed flying over the river while spraying pesticides on the surrounding banana plantations.

Figure 6. Site 5 (Desembocadura) corresponds to the mouth of the River Dos Novillos with the Parismina. The substrate consists almost entirely of sand. Big boulders are absent, small rocks are scarce and current flow is weak; tall grasses, banana plants, bamboo, and a few Agrorural Ecosystem Effects on the Macroinvertebrate Assemblages of a Tropical River 9 http://dx.doi.org/ 10.5772/59073 307

main criterion for this semi-quantitative collecting method is time; there were no defined sampling areas. All types of microhabitats present at a particular site were examined equally for the macroinvertebrates. Collected organisms were fixed immediately in 70 % ethanol at the time of sampling. Exact details of the sampling methodology can be found in Stein *et al.* [37]. **Figure 7.** Shallow waters and a moderate current characterize the tributary Quebrada Mercedes (site 6). Small rocks are the predominant substrate and current is moderate and laminar with some faster flowing riffle areas.

under normal current situations in order to avoid the negative effects of flooding and high water conditions.

substrate and current is moderate and laminar with some faster flowing riffle areas.

**2.2 Sampling**

Figure 7. Shallow waters and a moderate current characterize the tributary Quebrada Mercedes (site 6). Small rocks are the predominant

A plastic strainer with a diameter of 20 cm and 0.5 mm mesh size, and tweezers were used for directly collecting macroinvertebrates. The

A presampling was carried out in order to determine sampling time [37]. Out of the achieved results, an accumulated taxa curve was elaborated, and 120 min were determined to be a representative sampling time per site. Sampling took place during early morning and

The government of Costa Rica has officially suggested this method under the water quality monitoring regulation [38]. This method is coupled with the use of a modified Biological Monitoring Working Party index for for Costa Rica (BMWP´-CR), a biotic index utilized to

Figure 6. Site 5 (Desembocadura) corresponds to the mouth of the River Dos Novillos with the Parismina. The substrate consists almost entirely of sand. Big boulders are absent, small rocks are scarce and current flow is weak; tall grasses, banana plants, bamboo, and a few trees dominate vegetation along the channel. At this site, small airplanes were observed flying over the river while spraying pesticides on

the surrounding banana plantations. **Figure 6.** Site 5 (Desembocadura) corresponds to the mouth of the River Dos Novillos with the Parismina. The sub‐ strate consists almost entirely of sand. Big boulders are absent, small rocks are scarce and current flow is weak; tall grasses, banana plants, bamboo, and a few trees dominate vegetation along the channel. At this site, small airplanes

were observed flying over the river while spraying pesticides on the surrounding banana plantations.

3068 Biodiversity in Ecosystems - Linking Structure and Function


**Table 1.** Geographical and physical characteristics of each sampling site at the Dos Novillos River (1-5) and the Mercedes Stream (6), Guácimo, Province of Limón, Costa Rica.

### **2.2. Sampling**

A plastic strainer with a diameter of 20 cm and 0.5 mm mesh size, and tweezers were used for directly collecting macroinvertebrates. The main criterion for this semi-quantitative collecting method is time; there were no defined sampling areas. All types of microhabitats present at a particular site were examined equally for the macroinvertebrates. Collected organisms were fixed immediately in 70 % ethanol at the time of sampling. Exact details of the sampling methodology can be found in Stein *et al.* [37].

A presampling was carried out in order to determine sampling time [37]. Out of the achieved results, an accumulated taxa curve was elaborated, and 120 min were determined to be a representative sampling time per site. Sampling took place during early morning and under normal current situations in order to avoid the negative effects of flooding and high water conditions.

The government of Costa Rica has officially suggested this method under the water quality monitoring regulation [38]. This method is coupled with the use of a modified Biological Monitoring Working Party index for Costa Rica (BMWP´-CR), a biotic index utilized to define different levels of water quality. Each family of macroinvertebrates has a sensitivity value ranging from 1 to 10, reflecting tolerance to pollution based on the knowledge of distribution and abundance. The values for each family are then summed up independently from abun‐ dance and generic or species diversity. Sensitivity scores higher than 120 points indicate undisturbed aquatic ecosystems, while low values indicate serious contamination (mostly organic) of the environment [13, 23, 38,].

### **2.3. Data analysis**

The analyzed data comprised the values of the physical-chemical water quality variables (Table 2) and macroinvertebrate abundances (Table 3) during a collecting period of 13 months (January 2005 to January 2006).

The following physical-chemical variables were measured: pH, temperature, O2, O2 saturation, suspended solids, turbidity, conductivity, NO3 - , NH4 + , PO4 + , BOD, and COD. The water samples also were analyzed for the following agrochemicals associated with banana produc‐ tion using gas chromatography-MS and liquid chromatography-PDA: Chlorpyrifos, Diazinon, Dimethoate, Edifenfos, Etoprofos, Fenamifos, Malathion, Parathion-methyl, Parathion-ehtyl, Terbufos, Difenoconazol, Propiconazol, Imazalil, Ametrine, Atrazine, Hexazinone, Terbuty‐ lazine, Bromacil, Bitertanol, Chlorothalonil, and Thiabendazole. However, no traces of them could be detected in the river water samples. This does not come as a surprise because in order to monitor pesticides very frequent sampling would be required to detect peak concentrations during pesticide application periods [39], whereas low concentrations are very difficult to detect.

Following the suggestion of Ramírez and Gutiérrez-Fonseca [40], this study is also undertaking an ecosystem process analysis of the functional feeding groups (FGG) of the aquatic macro‐ invertebrates. This sort of analysis is based on two key aspects of macroinvertebrates: mor‐ phological characteristics related to the obtainment of food resources (*e.g.,* mouthparts and related structures) and behavioural mechanisms (*e.g.*, feeding behaviour). FGG is a very useful tool that provides valuable information on ecosystem functioning, facilitating stream ecosys‐ tem comparisons, and avoiding the traps of gut content analysis, which is more appropriate for assigning trophic guilds [40].

In this ecosystem study, the use of different parameters of the structure and composition of macroinvertebrate assemblages are presented: total and relative abundances, taxa richness, and functional feeding groups. Also, correlations of different genera and functional feeding groups with environmental variables were analyzed.

### **2.4. Statistical analysis**

**Site number 1 2 3 4 5 6**

Longitude (N) 10° 07´ 09.7´´ 10° 09´ 14.3´´ 10° 10´ 40.8´´ 10° 13´ 00.9´´ 10° 14´ 10° 12´ Latitude (W) 83° 39´ 15.2´´ 83° 37´ 24.7´´ 83° 36´ 10.6´´ 83° 35´ 18.4´´ 83° 34´ 83° 35´ Altitude (m) 441 187 90 51 40 44 Width (m) 24.2 17.5 30.5 24 22.2 9.5 Depth (m) 0.3-1.3 0.25-0.8 0.25-0.8 0.2-0.86 0.2 – 1.2 0.15-0.3 Current (m/s) 1.67 1.94 0.9 1.05 3.2 1.9

> Small rocks

A plastic strainer with a diameter of 20 cm and 0.5 mm mesh size, and tweezers were used for directly collecting macroinvertebrates. The main criterion for this semi-quantitative collecting method is time; there were no defined sampling areas. All types of microhabitats present at a particular site were examined equally for the macroinvertebrates. Collected organisms were fixed immediately in 70 % ethanol at the time of sampling. Exact details of the sampling

A presampling was carried out in order to determine sampling time [37]. Out of the achieved results, an accumulated taxa curve was elaborated, and 120 min were determined to be a representative sampling time per site. Sampling took place during early morning and under normal current situations in order to avoid the negative effects of flooding and high water

The government of Costa Rica has officially suggested this method under the water quality monitoring regulation [38]. This method is coupled with the use of a modified Biological Monitoring Working Party index for Costa Rica (BMWP´-CR), a biotic index utilized to define different levels of water quality. Each family of macroinvertebrates has a sensitivity value ranging from 1 to 10, reflecting tolerance to pollution based on the knowledge of distribution and abundance. The values for each family are then summed up independently from abun‐ dance and generic or species diversity. Sensitivity scores higher than 120 points indicate undisturbed aquatic ecosystems, while low values indicate serious contamination (mostly

The analyzed data comprised the values of the physical-chemical water quality variables (Table 2) and macroinvertebrate abundances (Table 3) during a collecting period of 13 months

**Table 1.** Geographical and physical characteristics of each sampling site at the Dos Novillos River (1-5) and the

**La Hamaca**

> Small rocks

**Desembocadura Quebrada**

Sand

**Mercedes**

Small rocks

**and name**

River bottom

**2.2. Sampling**

conditions.

**2.3. Data analysis**

**Don Eladio**

308 10 Biodiversity in Ecosystems - Linking Structure and Function

Medium-sized rocks

Mercedes Stream (6), Guácimo, Province of Limón, Costa Rica.

methodology can be found in Stein *et al.* [37].

organic) of the environment [13, 23, 38,].

(January 2005 to January 2006).

**La**

Medium-sized rocks

**Argentina Chiquitín**

The model comparisons between physical-chemical variables, macroinvertebrate abundances, and the BMWP´-CR index at different collecting sites was done by performing an analysis of variance (ANOVA; α=0.05). For all three cases, the proposed hypothesis is to test the existence of significant variable differences between sites. Abundances were square root transformed in order to comply with error normality. Evaluation of the best model (homocedastic or hetero‐ cedastic) for each variable was performed using the Akaike information criterion (AIC), which is one of the benchmarks of mixed models based on penalized likelihood [41, 42]. When the model detected significant differences, a DGC (Di Rienzo – González – Casanoves) statistical test was performed for the comparison of means [43].

On the other hand, taxonomic groupings and FFG relative frequencies analyses were done using a Chi-square test in order to assay for statistically significant differences between sites. The Chi-square analysis is testing for independence between the sites and the studied variables. Any p value below 0.05 shows the existence of an association between the site and the studied variable.

PLS regression is a technique that combines Principal Component Analysis and Linear Regression [44]. It is applied when it is desired to predict a set of dependent variables (y), in this case the abundance of macroinvertebrate genera and FFG abundances and the BMWP index values, from a set of predictor variables (x), in this case physical-chemical variables. To represent the results obtained from the PLS analysis, a Triplot graph was superimposed on a Biplot graph [45], thus correlating all variables. Then, the observations appear ordered in a Triplot graph (sites), depending on the values of the dependent variables (macroinvertebrate and FGG abundances and BMWP´-CR index) and their correlation with the predictor variables (physical-chemical water-quality variables). For the macroinvertebrate genera PLS analysis, out of the 127 collected taxa (of which, 123 could be identified to the genera level and their different developmental stages: larva, pupa, adult), the multimetric analysis included only 58 taxa, which were chosen using a PCA (Principal Component Analysis). The rest were charac‐ terized for repeating the same information. These 58 taxa, composed of 15 688 individuals, showed high projection values on the first two principal components. All statistical analyses were done using the InfoStat program [46].

In order to correlate the abundance of macroinvertebrates, FFG, and the BMWP´-CR index values of different sites, these variables were correlated with physical-chemical variables using the Spearman rank correlation coefficient. In the present case, the hypothesis tries to establish if one variable can be effectively substituted by another one, due to the existence of a significant correlation. The Spearman correlation coefficient was selected, versus Pearson, because its use is recommended in the case of having a small sample.

Finally, in order to evince the consistency of the sites´ congruences arranged in one plane unto physicochemical variables and FFG and macroinvertebrate genera abundances, a Generalized Procrustes Analysis was performed. This analysis is used for harmonizing multivariate configurations obtained on the same set of observations with different types of variables or time points [47]. Alignment is performed through a series of steps includ‐ ing normalization, rotation, reflection, and scaling of data to obtain a consensus array between groups of variables. This series of steps should maintain the distances between individuals from the individual configurations and minimize the distance between similar points [44]. The result of this multivariate method is to present a graph that displays the configurations arrived at by each variable type and the consensus configuration. A percentage consensus analysis was also undertaken. A high consensus indicates that any group of variables characterizes the different sites in the same way; therefore, using any group of variables is indistinct for site characterization.

## **3. Results**

### **3.1. Physical-chemical parameters**

Table 2 shows the results of the physical-chemical analysis (ANOVA, DGC-test, p>0, 05). All sites presented neutral pH and high dissolved oxygen levels and saturation. Temperatures varied in a statistically significant way, site 1 being the coolest place and site 5 the warmest. Conductivity was quite low at all sites, but statistically significantly lower in sites 5 and 6;


whereas, turbidity was statistically significantly higher in site 6. NO3 was statistically signifi‐ cantly higher in sites 5 and 6, and lowest in sites 1, 2, and 3.

this case the abundance of macroinvertebrate genera and FFG abundances and the BMWP index values, from a set of predictor variables (x), in this case physical-chemical variables. To represent the results obtained from the PLS analysis, a Triplot graph was superimposed on a Biplot graph [45], thus correlating all variables. Then, the observations appear ordered in a Triplot graph (sites), depending on the values of the dependent variables (macroinvertebrate and FGG abundances and BMWP´-CR index) and their correlation with the predictor variables (physical-chemical water-quality variables). For the macroinvertebrate genera PLS analysis, out of the 127 collected taxa (of which, 123 could be identified to the genera level and their different developmental stages: larva, pupa, adult), the multimetric analysis included only 58 taxa, which were chosen using a PCA (Principal Component Analysis). The rest were charac‐ terized for repeating the same information. These 58 taxa, composed of 15 688 individuals, showed high projection values on the first two principal components. All statistical analyses

In order to correlate the abundance of macroinvertebrates, FFG, and the BMWP´-CR index values of different sites, these variables were correlated with physical-chemical variables using the Spearman rank correlation coefficient. In the present case, the hypothesis tries to establish if one variable can be effectively substituted by another one, due to the existence of a significant correlation. The Spearman correlation coefficient was selected, versus Pearson, because its use

Finally, in order to evince the consistency of the sites´ congruences arranged in one plane unto physicochemical variables and FFG and macroinvertebrate genera abundances, a Generalized Procrustes Analysis was performed. This analysis is used for harmonizing multivariate configurations obtained on the same set of observations with different types of variables or time points [47]. Alignment is performed through a series of steps includ‐ ing normalization, rotation, reflection, and scaling of data to obtain a consensus array between groups of variables. This series of steps should maintain the distances between individuals from the individual configurations and minimize the distance between similar points [44]. The result of this multivariate method is to present a graph that displays the configurations arrived at by each variable type and the consensus configuration. A percentage consensus analysis was also undertaken. A high consensus indicates that any group of variables characterizes the different sites in the same way; therefore, using any

Table 2 shows the results of the physical-chemical analysis (ANOVA, DGC-test, p>0, 05). All sites presented neutral pH and high dissolved oxygen levels and saturation. Temperatures varied in a statistically significant way, site 1 being the coolest place and site 5 the warmest. Conductivity was quite low at all sites, but statistically significantly lower in sites 5 and 6;

were done using the InfoStat program [46].

310 12 Biodiversity in Ecosystems - Linking Structure and Function

is recommended in the case of having a small sample.

group of variables is indistinct for site characterization.

**3. Results**

**3.1. Physical-chemical parameters**

**Table 2.** Mean values of physical-chemical variables with their standard deviations used for the PLS analysis at the six collecting sites (January 2005-January 2006). Means in the same row with the same lettering are not significantly different (ANOVA, DGC-test, homocedastic model for the physical-chemical variables, p>0, 05).

### **3.2. Diversity and composition of macroinvertebrate assemblages**

The study collected a total of 17 163 specimens, distributed in the following 6 classes (number of total amount of specimens in parentheses): Clitellata (2), Turbellaria (10), Gastropoda (403), Arachnida (21), Malacostraca (40), and Insecta (16 706) with the following orders: Ephemer‐ optera (6299), Coleoptera (3150), Trichoptera (3010), Diptera (2868), Plecoptera (683), Odonata (435), Hemiptera (101), Megaloptera (91), Lepidoptera (64), Blattodea (2) (Table 3). The total abundance analysis (Figure 8) shows that Ephemeroptera, Coleoptera, Trichoptera, and Diptera were the most abundant groups, comprising 89.2 % of the collected macroinverte‐ brates.



Agrorural Ecosystem Effects on the Macroinvertebrate Assemblages of a Tropical River 15 http://dx.doi.org/ 10.5772/59073 313


**Class/Order Family Genus FFG**

312 14 Biodiversity in Ecosystems - Linking Structure and Function

**Site 1 Don Eladio**

Diptera Tabanidae Chrysops Pr 0 0 0 0 0 1

Diptera Tipulidae Hexatoma† Pr 5 2 3 23 26 82 141 Diptera Tipulidae Hexatoma-p Pr 0 0 3 0 0 2 5

**Site 2 La Argent.**

**Site 3 Chiquitín**

**Site 4 Hamaca**

**Site 5 Desembocadura**

**Site 6 Q.Merced**

**Total**



**Class/Order Family Genus FFG**

314 16 Biodiversity in Ecosystems - Linking Structure and Function

**Site 1 Don Eladio**

**Site 2 La Argent.**

**Site 3 Chiquitín**

**Site 4 Hamaca**

**Site 5 Desembocadura**

**Site 6 Q.Merced**

**Total**

**Table 3.** Total number of individuals collected (abundance) per genera and its different life-forms along the studied sites (a=adult, p=pupa, no sign=larva, † taxa considered for the genera PLS analysis). Functional feeding group (FFG) categories: CG=Collector-Gatherers, Ft=Filterers, Pc=Piercers, Pr=Predators, Sc=Scrapers, and Sh=Shredders.

**Figure 8.** Graph indicating the percentage taxonomic composition of the total study sample of all six collecting sites.

### **3.3. Comparison between sampling sites with different agrorural influence**

If we consider the mean total number of collected individuals per site, clear abundance differences are evinced (Figure 9). The greatest mean total numbers, with almost 300 individ‐ uals collected per sampling date, are found in the two least impacted sites (ANOVA, DGCtest; p>0.05),the reference (site 1) and the livestock-pasture site (site 2), whereas the sites under more intense human influence present statistically significantly reduced mean total values (around 150 individuals collected per month).

Table 4 presents a family, genera, and EPT richness analysis of the different collecting sites, as well as the mean BMWP-CR values and resulting water quality. The total family and genera richness does not show significant frequency differences between sampling sites, although the highest number of genera were found at the first two sites, with over 60 genera. The site with the lowest taxonomic richness, both on family and genus level, was site 5, close to the river mouth. The EPT taxa richness was highest at site 1 (14 EPT taxa), decreased downstream towards only nine EPT taxa at sites 5 and 6, although no statistically significant frequency differences were found using the Chi-square test. The BMWP'-CR index shows statistically significant differences (ANOVA, DGC-test; p>0.05), where the index values diminish accord‐ ing to the following site groupings: (sites 1-2)-(site 6)-(sites 3-4)-(site 5). The water quality, indicated by the mean BMWP'-CR index falls into the categories "good quality" (sites 1-2) and "regular quality" (sites 3-6).

**Figure 9.** Comparison of the mean of the total number of collected individuals in the six sampling sites. Means with the same letters are not significantly different (ANOVA, DGC-test; p>0.05).

On the other hand, a very clear change in the structure of macroinvertebrate assemblages can be observed along the different sampling sites of the Dos Novillos River (Figure 10). At the first site, the undisturbed sampling point, beetles are significantly more abundant than at the other sites (Chi-square, p<0.0001), which are all under the influence of human impact. The last


**Table 4.** Total number of families, genera richness, EPT richness (Ephemeroptera-Plecoptera-Trichoptera), and mean values for the BMWP'-CR index with resulting biological water quality in the different collecting sites. No statistically significant frequency differences were found using the Chi – square test for families, genera and EPT richness (ANOVA, DGC-test, heterocedastic model for the BMWP analysis, p>0, 05).

site (site 6) shows a significantly greater abundance of Gastropoda, Odonata, and Plecoptera (Chi-square, p<0.0001). Here the banana packing plant discharges effluents, carrying small banana pieces and other suspended organic material, into the water of the stream. Also, a tendency of an increase in mayfly (Ephemeroptera) abundance can be observed towards sites 4 and 5, while caddisfly abundance is relatively steady throughout all sampling sites.

**Figure 10.** Percentage abundances of the seven most common taxonomic groups at the six different sampling sites. Groups marked with an asterisk (\*) have statistically significant frequency differences (Chi-square; p<0.0001).

### **3.4. Functional feeding group analysis**

**3.3. Comparison between sampling sites with different agrorural influence**

(around 150 individuals collected per month).

316 18 Biodiversity in Ecosystems - Linking Structure and Function

*A A*

the same letters are not significantly different (ANOVA, DGC-test; p>0.05).

"regular quality" (sites 3-6).

0

50

100

150

Number of Individuals

200

250

300

350

If we consider the mean total number of collected individuals per site, clear abundance differences are evinced (Figure 9). The greatest mean total numbers, with almost 300 individ‐ uals collected per sampling date, are found in the two least impacted sites (ANOVA, DGCtest; p>0.05),the reference (site 1) and the livestock-pasture site (site 2), whereas the sites under more intense human influence present statistically significantly reduced mean total values

Table 4 presents a family, genera, and EPT richness analysis of the different collecting sites, as well as the mean BMWP-CR values and resulting water quality. The total family and genera richness does not show significant frequency differences between sampling sites, although the highest number of genera were found at the first two sites, with over 60 genera. The site with the lowest taxonomic richness, both on family and genus level, was site 5, close to the river mouth. The EPT taxa richness was highest at site 1 (14 EPT taxa), decreased downstream towards only nine EPT taxa at sites 5 and 6, although no statistically significant frequency differences were found using the Chi-square test. The BMWP'-CR index shows statistically significant differences (ANOVA, DGC-test; p>0.05), where the index values diminish accord‐ ing to the following site groupings: (sites 1-2)-(site 6)-(sites 3-4)-(site 5). The water quality, indicated by the mean BMWP'-CR index falls into the categories "good quality" (sites 1-2) and

> 1 2 3 4 5 6 Sites

**Figure 9.** Comparison of the mean of the total number of collected individuals in the six sampling sites. Means with

On the other hand, a very clear change in the structure of macroinvertebrate assemblages can be observed along the different sampling sites of the Dos Novillos River (Figure 10). At the first site, the undisturbed sampling point, beetles are significantly more abundant than at the other sites (Chi-square, p<0.0001), which are all under the influence of human impact. The last

*B*

*B*

*B*

*B*

The FFG analysis is presented in Figure 11 (Chi-square; p<0.0001). Collector-gatherers are the dominant group at all sites along the main river, with around 50% of all individuals collected; at Quebrada Mercedes (site 6), they are the second largest group after the filter-feeders. Filterfeeders show also high percentages at each site, with exception of site 1 (reference site), where they have statistically significant lower relative frequencies. One the other hand, piercers have a statistically significant greater relative frequency at this site, compared to the rest of the sampling sites. Finally, site 6 has a statistically significant lower relative frequency of collectorgatherers and greater relative frequencies of predators, scrapers, and shredders.

**Figure 11.** Percentage abundances of the functional feeding groups (FFG) at the six different sampling sites. Groups marked with an asterisk (\*) have statistically significant frequency differences (Chi-square; p<0.0001). CG=Collector-Gatherers, Ft=Filterers, Pc=Piercers, Pr=Predators, Sc=Scrapers, and Sh=Shredders.

The FFG abundance PLS (Figure 12) presents high variation explanatory values, with factor 1 explaining 39.1 % of the total variance and factor 2 explaining 25.4 %. The variables´ temper‐ ature, as well as the piercer and predator abundances, mainly separates the different sites on the horizontal projection, as does the shredder abundance on the vertical projection. The FFG most closely allied to the reference site, is piercer abundance. On the other hand, the most closely allied variables to the agriculturally impacted areas, sites 5 and 6, are the variables turbidity, conductivity, and NO3.

### **3.5. Relationship between macroinvertebrate assemblages, physical-chemical parameters and sampling sites**

The macroinvertebrate abundance partial least square analysis (Figure 13) presents high variation explanatory values, with factor 1 explaining 42.0 % of the total variance and factor 2 explaining 29.9%. The variable NO3, and to a lesser degree conductivity and turbidity, separate mainly the different sites on the horizontal projection, as does suspended solids on the vertical

Agrorural Ecosystem Effects on the Macroinvertebrate Assemblages of a Tropical River 21 http://dx.doi.org/ 10.5772/59073 319

feeders show also high percentages at each site, with exception of site 1 (reference site), where they have statistically significant lower relative frequencies. One the other hand, piercers have a statistically significant greater relative frequency at this site, compared to the rest of the sampling sites. Finally, site 6 has a statistically significant lower relative frequency of collector-

CG Ft Pc Pr Sc Sh

**Figure 11.** Percentage abundances of the functional feeding groups (FFG) at the six different sampling sites. Groups marked with an asterisk (\*) have statistically significant frequency differences (Chi-square; p<0.0001). CG=Collector-

The FFG abundance PLS (Figure 12) presents high variation explanatory values, with factor 1 explaining 39.1 % of the total variance and factor 2 explaining 25.4 %. The variables´ temper‐ ature, as well as the piercer and predator abundances, mainly separates the different sites on the horizontal projection, as does the shredder abundance on the vertical projection. The FFG most closely allied to the reference site, is piercer abundance. On the other hand, the most closely allied variables to the agriculturally impacted areas, sites 5 and 6, are the variables

**3.5. Relationship between macroinvertebrate assemblages, physical-chemical parameters**

The macroinvertebrate abundance partial least square analysis (Figure 13) presents high variation explanatory values, with factor 1 explaining 42.0 % of the total variance and factor 2 explaining 29.9%. The variable NO3, and to a lesser degree conductivity and turbidity, separate mainly the different sites on the horizontal projection, as does suspended solids on the vertical

Gatherers, Ft=Filterers, Pc=Piercers, Pr=Predators, Sc=Scrapers, and Sh=Shredders.

1 2 3 4 5 6 Site

*\**

Sh

*\**

*\**

*\**

turbidity, conductivity, and NO3.

**and sampling sites**

Individuals (%)

*\**

*\**

318 20 Biodiversity in Ecosystems - Linking Structure and Function

gatherers and greater relative frequencies of predators, scrapers, and shredders.

**Figure 12.** Site ordination (1-6) with a triplot PLS analysis using physical-chemical values as independent variables and functional feeding group abundances as dependent variables. CG=Collector-Gatherers, Ft=Filterers, Pc=Piercers, Pr=Predators, Sc=Scrapers, and Sh=Shredders.

one. The genera most closely allied to the reference site are: *Brysopterix*, *Heterelmis*, and *Paltostoma*. On the other hand, the most closely allied genera to the agriculturally impacted areas, sites 5 and 6, are larvae of the genera *Caenis*, *Dythemis, and Hexatoma*. Interestingly, the Spearman rank analysis has resulted in a plethora of correlations. The strongest physicalchemical/biological correlations are presented in Table 5.



**Table 5.** Statistically significant associations of the Spearman rank correlation coefficient between physical-chemical, genera, and functional feeding groups (a=adults, p=pupae, genera with no lettering=larvae).

**Figure 13.** Site ordination (1-6) with a triplot PLS analysis, using physical-chemical values as independent variables, and macroinvertebrate generic abundance and the BMWP index as dependent variables (FFG categories: purple=col‐ lector-gatherers, yellow=filterers, green=piercers, grey=predators, red=scrapers, blue=shredders).

The Procrustes analysis produced a good site ordination consensus based on macroinverte‐ brate abundances, FFG, and the physical-chemical variables. The first axis explains 45.8 % of the variance and the second axis explains 33.8 % (Figure 14). The proportional consensus percentages are also good, ranging from 76 % to 92.4 %, with a mean value of 82.9% (Table 6). One can conclude that the site ordination has a good congruence with the biological and physical-chemical data sets.

Agrorural Ecosystem Effects on the Macroinvertebrate Assemblages of a Tropical River 23 http://dx.doi.org/ 10.5772/59073 321

**Figure 14.** Site ordination configuration congruence according to physical-chemical values (PC), macroinvertebrate and functional feeding group (FFG) abundances (Genera) using a Procrustes analysis.


**Table 6.** Proportional consensus percentages as displayed by the configurations generated between macroinvertebrate generic abundances, physical-chemical values (PC), and functional feeding groups (FFG) with the generalized Procrustes ordination.

### **4. Discussion**

**Variable (1) Variable (2) Spearman rank p-value** Turbidity Cora -0.93 0.0080 PO4 *Cylloepus*-a -0.93 0.0080 PO4 *Macrelmis*-a -0.93 0.0080 PO4 pH 0.93 0.0080 NO3 PO4 0.93 0.0080

**Table 5.** Statistically significant associations of the Spearman rank correlation coefficient between physical-chemical,

*Chloronia*

*Tricorythodes*

*BMWP*


**Figure 13.** Site ordination (1-6) with a triplot PLS analysis, using physical-chemical values as independent variables, and macroinvertebrate generic abundance and the BMWP index as dependent variables (FFG categories: purple=col‐

The Procrustes analysis produced a good site ordination consensus based on macroinverte‐ brate abundances, FFG, and the physical-chemical variables. The first axis explains 45.8 % of the variance and the second axis explains 33.8 % (Figure 14). The proportional consensus percentages are also good, ranging from 76 % to 92.4 %, with a mean value of 82.9% (Table 6). One can conclude that the site ordination has a good congruence with the biological and

*Chimarra*

**O2 Saturation**

**5 Turbidity 6**

*Nehalenia*

*Epiphrades Farrodes*

**NO3**

**PO4**

*Cylloepus*

*Caenis*

**Temperature**

**pH**

**COD**

**Conductivity**

*Dythemis*

**NH4**

*Triplectides*

*Hexanchorus Hexatoma*

*Heteragrion*

**Solids**

*Paltostoma-p Perrisolestes*

**3 4**

*Simulium*

*Macrothemis*

*Atopsyche-p*

*Leptohyphes*

lector-gatherers, yellow=filterers, green=piercers, grey=predators, red=scrapers, blue=shredders).





0

Factor 2 (29.9%)

1

2

3

4

genera, and functional feeding groups (a=adults, p=pupae, genera with no lettering=larvae).

*Phanocerus-a*

*Asioplax*

*Cochliopsyche*

*Maruina-p*

*Brechmorhoga*

*Oxyethira*

*Atanatolica-p*

*Paltostoma*

*Brysopterix*

*Phanocerus*

**1**

*Mayobaetis*

*Atanatolica*

*Heterelmis-a Hexanchorus-a*

320 22 Biodiversity in Ecosystems - Linking Structure and Function

**2**

*Neoelmis-a*

*Microcylloepus-a*

*Hebrus*

*Macrelmis*

*Cora*

*Psephenus*

*Austrolimnius-a Baetodes*

*Macrelmis-a*

*Cylloepus-a*

*Maruina*

*Camelobaetidius*

physical-chemical data sets.

*Petrophila*

*Thraulodes*

*Brysopterix-p*

*Ochterus*

*Desmogomphus*

*Heterelmis*

*Polycentropus*

*Pseudodisersus*

**O2**

*Hemerodromia*

*Anchytarsus*

*Atopsyche*

**BOD**

An important consideration for the present study is the ability to distinguish between natural variability and human impacts [49]. In the present case, all sites are located along the same river, spanning only a small distance of 10 km (Figure 1). Moreover, elevation, stream size, and surface geology are relatively similar (Table 1) in order to properly assess human impacts and reduce natural gradients.

### **4.1. Diversity and composition of macroinvertebrate assemblages**

The present study recorded the existence of 16 macroinvertebrate orders. In an analysis undertaken by Castillo *et al*. [17] very near to the present collecting localities, the authors reported the existence of 15-16 macroinvertebrate orders in their reference sites and 12-16 orders in the banana plantation sites. The results of the present study fall within the order range for similar studies in nearby regions. The present analysis resulted in 53 collected families (Table 3). This number compares well with the number of families collected by Lorion and Kennedy [27] and O'Callaghan and Kelly-Quinn [58] in Costa Rican and Honduran neotropical rivers, who reported 56 and 60 families, respectively. The present study also identified a total of 98 genera (Table 3). Montoya Moreno *et al*. [59] reported 69 genera for the Negro River in Colombia; whereas, Sánchez Argüello *et al*. [59] reported 96 genera in their study in Panama. Considering that some groups, such as water mites, were not identified to genus level in the present study, the total number of genera present in the Dos Novillos river and its tributaries is likely to be over 100 genera, reflecting a very high taxa richness.

It is interesting to compare the dominance results obtained in this study with analyses made in neighbouring areas under similar ecological conditions. Ramírez *et al*. [33] sampled the Carbón and Gandoca Rivers and found the following dominance gradient for total abundan‐ ces: Ephemeroptera-Diptera-Trichoptera-Odonata. Castillo *et al*. [17], sampling on sites very near to the present ones, found the following order dominance in their reference sites: Ephemeroptera-Trichoptera-Coleoptera and the following one in banana plantation sites: Ephemeroptera-Diptera-Coleoptera-Gastropoda-Trichoptera. Lorion and Kennedy [27] studied several streams in the Sixaola River Valley. Considering the total number of individ‐ uals, they found a diminishing abundance gradient as follows: Ephemeroptera-Diptera-Coleoptera-Odonata. Gutiérrez-Fonseca and Ramírez [50] have reported at La Selva Biological Station, in a 15-year study, the following dominance sequence in unpolluted streams: Diptera-Trichoptera-Odonata. The present study (Figure 8) has found the following abundance dominance order: Ephemeroptera-Coleoptera-Trichoptera-Diptera. It would seem from these results that dominance sequences might vary depending on several factors dependent on the collecting site, such as substrate, current, and water quality, but also can be a result of different sampling device and mesh size [37]. However, Ephemeroptera would appear to be the most constant dominant group in most lowland rivers and streams in Costa Rica.

### **4.2. Comparison between sampling sites with different agrorural influence**

The analysis of abundance differences demonstrates that less impacted sites clearly show statistically significantly higher abundances than sites under stronger human influence (Figure 9). Similarly, Paaby *et al*. [28] and Lorion and Kennedy [27] detected greater abundances in forested areas versus pastures under neotropical conditions; whereas, the study by Ramírez *et al*. [33] detects greater abundances in Costa Rican tropical riffle habitats than in other habitats. Taxonomic composition also varies in a clear way along the sampling sites under the influence of different ecological impacts (Figure 10). The reference site, arguably not under the influence of human impact, presented a significantly greater relative abundance of Coleoptera. This is due to the high amount of individuals collected from the riffle beetle family (Elmidae), which are especially diverse and abundant in well oxygenized rivers and streams in forested areas (Springer, *unpubl*). A study of rivers in the Guanacaste area in Costa Rica (Kohlmann, *unpubl.*) has also found very high numbers of Coleoptera in unpolluted rivers. At the other end of the scale, the banana packing plant discharge, carrying banana debris and showing statistically significantly higher values of conductivity, turbidity, and NO3, presents statisti‐ cally significantly high numbers of Gastropoda, Odonata, and Plecoptera. This is very interesting because stoneflies (Plecoptera) have always been considered as good indicators of oxygenated, clean, and cool running waters [51-54]. Additionally, the triplot (Figure 12) and Spearman rank correlation analyses show a strong negative correlation between stonefly (*Anacroneuria*) abundance and temperature values (Spearman rank correlation value=-0.89, p=0.0476). Gutiérrez and Springer [55] reported the widespread species *A. holzenthali* from coffee plantations in Costa Rica. Tomanova and Tedesco [56], as well as Thorp and Covich [57] also indicated that stonefly presence is not necessarily a sure sign of water cleanliness. The results from these analyses seem to support this claim.

and surface geology are relatively similar (Table 1) in order to properly assess human impacts

The present study recorded the existence of 16 macroinvertebrate orders. In an analysis undertaken by Castillo *et al*. [17] very near to the present collecting localities, the authors reported the existence of 15-16 macroinvertebrate orders in their reference sites and 12-16 orders in the banana plantation sites. The results of the present study fall within the order range for similar studies in nearby regions. The present analysis resulted in 53 collected families (Table 3). This number compares well with the number of families collected by Lorion and Kennedy [27] and O'Callaghan and Kelly-Quinn [58] in Costa Rican and Honduran neotropical rivers, who reported 56 and 60 families, respectively. The present study also identified a total of 98 genera (Table 3). Montoya Moreno *et al*. [59] reported 69 genera for the Negro River in Colombia; whereas, Sánchez Argüello *et al*. [59] reported 96 genera in their study in Panama. Considering that some groups, such as water mites, were not identified to genus level in the present study, the total number of genera present in the Dos Novillos river

and its tributaries is likely to be over 100 genera, reflecting a very high taxa richness.

constant dominant group in most lowland rivers and streams in Costa Rica.

**4.2. Comparison between sampling sites with different agrorural influence**

The analysis of abundance differences demonstrates that less impacted sites clearly show statistically significantly higher abundances than sites under stronger human influence (Figure 9). Similarly, Paaby *et al*. [28] and Lorion and Kennedy [27] detected greater abundances in forested areas versus pastures under neotropical conditions; whereas, the study by Ramírez *et al*. [33] detects greater abundances in Costa Rican tropical riffle habitats than in other habitats.

It is interesting to compare the dominance results obtained in this study with analyses made in neighbouring areas under similar ecological conditions. Ramírez *et al*. [33] sampled the Carbón and Gandoca Rivers and found the following dominance gradient for total abundan‐ ces: Ephemeroptera-Diptera-Trichoptera-Odonata. Castillo *et al*. [17], sampling on sites very near to the present ones, found the following order dominance in their reference sites: Ephemeroptera-Trichoptera-Coleoptera and the following one in banana plantation sites: Ephemeroptera-Diptera-Coleoptera-Gastropoda-Trichoptera. Lorion and Kennedy [27] studied several streams in the Sixaola River Valley. Considering the total number of individ‐ uals, they found a diminishing abundance gradient as follows: Ephemeroptera-Diptera-Coleoptera-Odonata. Gutiérrez-Fonseca and Ramírez [50] have reported at La Selva Biological Station, in a 15-year study, the following dominance sequence in unpolluted streams: Diptera-Trichoptera-Odonata. The present study (Figure 8) has found the following abundance dominance order: Ephemeroptera-Coleoptera-Trichoptera-Diptera. It would seem from these results that dominance sequences might vary depending on several factors dependent on the collecting site, such as substrate, current, and water quality, but also can be a result of different sampling device and mesh size [37]. However, Ephemeroptera would appear to be the most

**4.1. Diversity and composition of macroinvertebrate assemblages**

and reduce natural gradients.

322 24 Biodiversity in Ecosystems - Linking Structure and Function

The EPT taxa richness (Table 4) showed a steady decline along the collecting sites, following an increasing anthropogenic impact trend, similar to what Lorion and Kennedy [27] reported following a forest-forest buffer-pasture gradient. The number of families per site (Table 4) did not vary much (33-40) along the collecting transect in this study (Table 4). An analysis by Castillo *et al*. [17] reported a family number variation going from 39 to 47 families. Similarly, their family numbers did not vary much between their reference (46-47) and their banana plantation collecting sites (39-46). Some families (Table 3) were restricted to only a single site (site number in parentheses), like (1): Anomalopsychidae, Calamoceratidae; (3): Ancylidae, Lumbriculidae, Physidae; (4): Corduliidae; (5): Gyrinidae; (6): Ampullaridae, Chironomidae, and Tabanidae. Certain genera (Table 3) appeared only once in a specific site: (1): *Chloronia, Contulma, Dythemis, Hemerodromia, Ochterus, Palaemnema,* and *Polycentropus*; (2): *Hebrus, Leptohyphes, Mayobaetis,* and *Psephenops*; (3) *Stenonema*; (4): *Anchitrichia* and *Neocylloepus*; (5): *Haplohyphes*; and (6): *Chironomus, Chrysops*. These taxa would seem to be closely associated with the ecological conditions of each site, *e.g.* Ampullaridae, Chironomidae, and Tabanidae in a banana packing-plant effluent with a high organic waste discharge versus Anomalopsy‐ chidae and Calamoceratidae in a forested undisturbed condition. Interestingly, the number of families and genera do not show the same trend among sampling sites. The reference and slightly disturbed sites (site 1 and 2, respectively) show a higher number of genera (over 60), in comparison to the more influenced sites (with around 40 to 50 genera), even though the number of families were similar at the latter, and even higher in one case (site 3). These results suggest that family richness is not necessarily an adequate indicator for biomonitoring, and generic identification is necessary to achieve results that are more reliable.

The BMWP index in its different variations has been popularly employed in Latin America. These studies usually have found this index to be satisfactory for reflecting water quality [10, 12, 22, 60-63], especially the Costa Rican adaptation [58]. Sánchez Argüello *et al*. [61] undertook in their Panamanian study a comparison between the Colombian and Costa Rican adaptations of the BMWP index and found the latter to be more unforgiving in its water quality evaluation. Rizo-Patrón *et al*. [22] undertook an analysis using the BMWP index modified for Costa Rica, studying the environmental impact caused by conventional and organic-irrigated rice fields on the macroinvertebrate communities. Their BMWP´-CR results show that the index values were greater in the organically irrigated rice fields. On the other hand, Fenoglio *et al*. [64] recommend, from their experience in Nicaragua, the use of the Indice Biotico Esteso [65] because of its ease of use and low cost. However, these studies have mostly assessed the comparative performance of the various indices; no attempts were made to correlate the BMWP index to specific physical-chemical variables.

The results of the BMWP'-CR index for the present analysis (Table 2) do indeed show a discriminating capacity of the index following a diminishing environmental quality site trend, especially under agricultural impact conditions (site 5), but it also shows a tendency of reporting a higher value when in river waters with high organic pollution (site 6). Interestingly, there is a statistically significant negative correlation of the BMWP'-CR index with tempera‐ ture, not with pollution variables, as one could expect from the general assumption that the BMWP index reflects organic water pollution quality. These results generate some doubts about the reliability of the BMWP'-CR index as an environmentally representative tool, as the following studies indicate. Sermeño Chicas *et al*. [66] tried to implement the BMWP-CR index in El Salvador where rivers showed consistently high organic pollution conditions that were not reflected by the BMWP index [67]. In their analysis of selected macroinvertebrate-based biotic indices in Honduras, O´Callaghan and Kelly-Quinn [58] found that a BMWP-CR-based version of the ASPT index performed much better than the aforementioned index. Without doubt, more studies will be necessary in order to adjust the biotic indices used for aquatic biomonitoring in Costa Rica and Central America according to the different ecoregions.

### **4.3. Functional feeding group analysis**

FFG relative abundances also change significantly depending on the human impact conditions on the quality of river water. It would seem that under undisturbed conditions filterers´ relative abundances tend to be minimal, their increase at disturbed sites might be a result of higher dissolved organic matter. In this study, under conditions of high organic pollution, shredders, scrapers, and predators tend to have maximal relative abundances while collector-gatherers tend to have minimal values. Finally, filter feeders seem to react positively to high concentra‐ tions of dissolved O2 (Table 5), which is positively correlated with fast flowing waters, a condition that also favors the feeding mechanism of filterers.

The taxonomic grouping triplot analysis (Figure 13) suggests a correlation between the reference site and the piercers, and it would appear that the first axis is characterized by a piercers' abundance gradient, diminishing from the reference community (site 1) to the high organic waste discharge sites (site 6). In the present study, piercers are mainly represented by the caddisfly family Hydroptilidae, which is especially abundant in the splash zone of big rocks in riffle areas, which were characteristic for the reference sampling site. The second axis appears to be characterized by filterers, arranging the different agrorural ecosystems along a diminishing abundance gradient.

The FFG triplot analysis (Figure 12) supports a strong agrorural ecosystem ordination process mediated by the abundance of piercers along the main axis as suggested in the previous triplot analysis; however, as this analysis benefits from having more information (127 taxa versus 58 taxa), it also evinces the importance of predators and shredders as relevant ecosystem ordinating biological variables. The strong negative correlation between the piercer's abun‐ dance and NO3 and PO4 values (Table 5) stresses again the importance of the piercers as an ecosystem characterizing variable, although the presence of suitable microhabitats might be another important factor to consider.

### **4.4. Relationship between macroinvertebrate assemblages, physical-chemical parameters and sampling sites**

The taxonomic grouping triplot analysis (Figure 13) and the Spearman rank correlation analysis (Table 5) show the existence of several genera and FFG that are highly correlated with physical-chemical variables and that possibly could be used as surrogates for these variables. The larvae of *Maruina* showed one of the strongest correlations with NO3, although other genera like *Farrodes* and *Cora,* and the piercers´ functional feeding group were also significantly correlated with this chemical variable. Species of the genus *Caenis* have been found regularly in organically enriched streams [48]. Of the statistically significant genus list (Table 5), only *Heterelmis* and *Farrodes* already had been cited before as good quality bioindicators for toxicity and pollution-sensitivity testing by Castillo *et al.* [17] and Rizo-Patrón *et al.* [22], respectively, in a similar type of analysis.

Finally, the Procrustes analysis allows an assessment of the goodness of fit of the taxonomic, FFG, and physical-chemical analyses (Figure 14, Table 6). The analysis resulted in a good match of the collecting sites (landmarks) with the values derived from the three blocks of variables, indicating that any one of the three is describing in a similar way the ecology of each study site. The consensus values indicated in Table 6 show a very good concordance in this study between the different variable blocs and the consensus values, where generic abundances seem to generate a better ecological ordination. An environmental impact gradient also becomes very apparent on the first ordination axis, ranging from the undisturbed reference site inside the forest on the right of the graph to the banana packing-plant effluent on the left.

## **5. Conclusions**

12, 22, 60-63], especially the Costa Rican adaptation [58]. Sánchez Argüello *et al*. [61] undertook in their Panamanian study a comparison between the Colombian and Costa Rican adaptations of the BMWP index and found the latter to be more unforgiving in its water quality evaluation. Rizo-Patrón *et al*. [22] undertook an analysis using the BMWP index modified for Costa Rica, studying the environmental impact caused by conventional and organic-irrigated rice fields on the macroinvertebrate communities. Their BMWP´-CR results show that the index values were greater in the organically irrigated rice fields. On the other hand, Fenoglio *et al*. [64] recommend, from their experience in Nicaragua, the use of the Indice Biotico Esteso [65] because of its ease of use and low cost. However, these studies have mostly assessed the comparative performance of the various indices; no attempts were made to correlate the

The results of the BMWP'-CR index for the present analysis (Table 2) do indeed show a discriminating capacity of the index following a diminishing environmental quality site trend, especially under agricultural impact conditions (site 5), but it also shows a tendency of reporting a higher value when in river waters with high organic pollution (site 6). Interestingly, there is a statistically significant negative correlation of the BMWP'-CR index with tempera‐ ture, not with pollution variables, as one could expect from the general assumption that the BMWP index reflects organic water pollution quality. These results generate some doubts about the reliability of the BMWP'-CR index as an environmentally representative tool, as the following studies indicate. Sermeño Chicas *et al*. [66] tried to implement the BMWP-CR index in El Salvador where rivers showed consistently high organic pollution conditions that were not reflected by the BMWP index [67]. In their analysis of selected macroinvertebrate-based biotic indices in Honduras, O´Callaghan and Kelly-Quinn [58] found that a BMWP-CR-based version of the ASPT index performed much better than the aforementioned index. Without doubt, more studies will be necessary in order to adjust the biotic indices used for aquatic biomonitoring in Costa Rica and Central America according to the different ecoregions.

FFG relative abundances also change significantly depending on the human impact conditions on the quality of river water. It would seem that under undisturbed conditions filterers´ relative abundances tend to be minimal, their increase at disturbed sites might be a result of higher dissolved organic matter. In this study, under conditions of high organic pollution, shredders, scrapers, and predators tend to have maximal relative abundances while collector-gatherers tend to have minimal values. Finally, filter feeders seem to react positively to high concentra‐ tions of dissolved O2 (Table 5), which is positively correlated with fast flowing waters, a

The taxonomic grouping triplot analysis (Figure 13) suggests a correlation between the reference site and the piercers, and it would appear that the first axis is characterized by a piercers' abundance gradient, diminishing from the reference community (site 1) to the high organic waste discharge sites (site 6). In the present study, piercers are mainly represented by the caddisfly family Hydroptilidae, which is especially abundant in the splash zone of big rocks in riffle areas, which were characteristic for the reference sampling site. The second axis

BMWP index to specific physical-chemical variables.

324 26 Biodiversity in Ecosystems - Linking Structure and Function

**4.3. Functional feeding group analysis**

condition that also favors the feeding mechanism of filterers.

The present study clearly shows that tropical river macroinvertebrate diversity changes and at the same time characterizes and defines different river ecosystem conditions under various agrorural impacts. Changes in taxonomic composition and functional feeding group structure are very indicative of ecosystem function.

Ephemeroptera seem to be, in general, a rather constant and numerous group, present in the great majority of collections in neotropical rivers. However, high relative abundances of Coleoptera, especially from the family Elmidae, seem to be indicative of unpolluted conditions in tropical rivers; whereas, high relative abundances of Gastropoda, Odonata, and Plecoptera show up in sites with relatively high (plant-derived) organic pollution, although with well-oxygenized waters and forested stream margins. Addtionally, high numbers of individuals were found in unpolluted or slightly polluted sites; whereas, lower abundances were found in sites under human impact (town and fruit packing plant discharges, agricultural plantations).

Especially illustrative is the change in structure of functional feeding groups. Piercers showed the highest relative abundance in unpolluted sites and seem to be especially sensitive to human impacts because they quickly disappear under altered conditions, most probably reacting to the decrease of their microhabitat and food items. Filterers have the lowest relative abundance under unpolluted conditions and quickly become relatively more abundant under human impact conditions, most probably reflecting an increase in particles in the water column. At the other end of the spectrum, under high (plant-derived) organic pollution, shredders and scrapers show their highest relative abundance concomitantly with an increase in particulate organic matter, and probably as a response to it. Predators here show their highest relative abundance as well and at the same time, collector-gatherers here show their lowest relative abundance. It would seem then that predation, as well as scraping and shredding, increases significantly in river areas with high (plant-derived) organic pollution.

The biomonitoring analysis presented in this chapter used an adaptation of the BMWP index to Costa Rica. In all cases, the method revealed or indicated the existence of an anthropogenic/ agricultural impact gradient going from the unpolluted site to the most perturbed locality. ANOVA tests evidenced the fact that the BMWP index has enough sensitivity and discrimi‐ nating power to detect changes in macroinvertebrate biodiversity, which can be translated into statistically significant differences between sampling sites. The method determined the existence of changes in macroinvertebrate communities associated with agricultural areas, even when analytical methods could not detect the presence of pesticides in river water. A previous analysis of the BMWP score by Pinder and Farr [68, 69] found that it was significantly negatively correlated only with dissolved organic carbon. The present study detects a strong negative correlation between the BMWP score and temperature (Figure 12, Table 5). Due to its simplicity, speed of use, efficiency, and cheapness, this method shall undoubtedly continue to be a very popular one in the future. The BMWP is considered an extremely successful index according to Spellerberg [70]. However, there are some doubts about its suitability as a tool for detecting organic pollution in some regions of Latin America, especially if one considers that in this case the index was more sensitive to the impact of a banana plantation than the one caused by relatively high (plant-derived) organic pollution. The PLS/Procrustes analysis seemed, on this occasion, to be a more suitable method for describing and evaluating anthro‐ pogenic/agricultural environmental impacts.

The use of multivariate ordination for environmental studies is becoming more and more common. In particular, the PLS/Procrustes analyses represent a very powerful combination tool as they not only perform a site ordination but different taxa and environmental variables can be correlated at the same time. This makes this method extremely useful for taxa, physicalchemical, and FFG variables' correlations as specific environmental variable surrogates. Castillo *et al*. [17] and Rizo-Patrón *et al*. [22] found similar promising results for the study of agricultural ecosystem analyses. Castillo *et al*. [17] indicated, based on their results, that multivariate analyses are more sensitive in distinguishing pesticide effects than toxicity tests. Therefore, multivariate analyses should be incorporated as an approach for future ecosystem/ biodiversity analyses.

## **Acknowledgements**

of Coleoptera, especially from the family Elmidae, seem to be indicative of unpolluted conditions in tropical rivers; whereas, high relative abundances of Gastropoda, Odonata, and Plecoptera show up in sites with relatively high (plant-derived) organic pollution, although with well-oxygenized waters and forested stream margins. Addtionally, high numbers of individuals were found in unpolluted or slightly polluted sites; whereas, lower abundances were found in sites under human impact (town and fruit packing plant

Especially illustrative is the change in structure of functional feeding groups. Piercers showed the highest relative abundance in unpolluted sites and seem to be especially sensitive to human impacts because they quickly disappear under altered conditions, most probably reacting to the decrease of their microhabitat and food items. Filterers have the lowest relative abundance under unpolluted conditions and quickly become relatively more abundant under human impact conditions, most probably reflecting an increase in particles in the water column. At the other end of the spectrum, under high (plant-derived) organic pollution, shredders and scrapers show their highest relative abundance concomitantly with an increase in particulate organic matter, and probably as a response to it. Predators here show their highest relative abundance as well and at the same time, collector-gatherers here show their lowest relative abundance. It would seem then that predation, as well as scraping and shredding, increases

The biomonitoring analysis presented in this chapter used an adaptation of the BMWP index to Costa Rica. In all cases, the method revealed or indicated the existence of an anthropogenic/ agricultural impact gradient going from the unpolluted site to the most perturbed locality. ANOVA tests evidenced the fact that the BMWP index has enough sensitivity and discrimi‐ nating power to detect changes in macroinvertebrate biodiversity, which can be translated into statistically significant differences between sampling sites. The method determined the existence of changes in macroinvertebrate communities associated with agricultural areas, even when analytical methods could not detect the presence of pesticides in river water. A previous analysis of the BMWP score by Pinder and Farr [68, 69] found that it was significantly negatively correlated only with dissolved organic carbon. The present study detects a strong negative correlation between the BMWP score and temperature (Figure 12, Table 5). Due to its simplicity, speed of use, efficiency, and cheapness, this method shall undoubtedly continue to be a very popular one in the future. The BMWP is considered an extremely successful index according to Spellerberg [70]. However, there are some doubts about its suitability as a tool for detecting organic pollution in some regions of Latin America, especially if one considers that in this case the index was more sensitive to the impact of a banana plantation than the one caused by relatively high (plant-derived) organic pollution. The PLS/Procrustes analysis seemed, on this occasion, to be a more suitable method for describing and evaluating anthro‐

The use of multivariate ordination for environmental studies is becoming more and more common. In particular, the PLS/Procrustes analyses represent a very powerful combination tool as they not only perform a site ordination but different taxa and environmental variables can be correlated at the same time. This makes this method extremely useful for taxa, physical-

significantly in river areas with high (plant-derived) organic pollution.

discharges, agricultural plantations).

326 28 Biodiversity in Ecosystems - Linking Structure and Function

pogenic/agricultural environmental impacts.

We would like to thank Mildred Linkimer for her help in the design of some figures; as well as Roxana Araya for her help in the composition of tables and lists. Finally yet importantly, we are also indebted to Jane Yeomans and Kent McLeod for critically revising the text and to Pablo Gutiérrez for useful suggestions in organizing the results. This material is based upon work supported by the Department of Energy [National Nuclear Security Administration] under Award Number DE FG02-04ER 63856.

## **Author details**

Bert Kohlmann1\*, Alejandra Arroyo1 , Monika Springer2 and Danny Vásquez3


## **References**


[17] Castillo E., Martínez E., Ruepert C., Savage C., Gilek M., Pinnock M., Solís E. Water Quality and Macroinvertebrate Community Response Following Pesticide Applica‐ tions in a Banana Plantation, Limón, Costa Rica. Science of the Total Environment 2006; 367: 418-432.

[4] Kohlmann B, Mitsch WJ, Hansen DO. Ecological Management and Sustainable De‐ velopment in the Humid Tropics of Costa Rica. Ecological Engineering 2008; 34(4):

[5] Jones JR. Colonization and Environment: Land Settlement Projects in Central Ameri‐

[6] Seligson M. Peasants of Costa Rica and the Development of Agrarian Capitalism.

[7] de Oñoro MT. El Asentamiento Neguev. Atlantic Zone Programme. Inf. Técnico No. 162, Programme Paper No. 7. Turrialba: Centro Agronómico Tropical de Investiga‐ ción y Enseñanza; Universidad Agrícola de Wageningen; Ministerio de Agricultura y

[8] Roldán G., Builes J., Trujillo CM., Suárez A. Efectos de la Contaminación Industrial y Doméstica sobre la Fauna Béntica del Río Medellín. Actualidades Biológicas 1973;

[9] Roldán G. Los Macroinvertebrados y su Valor como Indicadores de la Calidad del Agua. Revista de la Academia Colombiana de Ciencias Exactas, Físicas y Naturales

[10] Roldán G. Bioindicación de la Calidad del Agua en Colombia. Antioquia: Universi‐

[11] de la Lanza Espino G. Criterios Generales para la Elección de Bioindicadores. In: de la Lanza Espino G., Hernández Pulido H., Carbajal Pérez JL. (eds.). Organismos Indi‐ cadores de la Calidad del Agua y de la Contaminación (Bioindicadores). México,

[12] Prat N., Ríos B., Acosta R., Rieradevall M. Los Macroinvertebrados como Indicadores de la Calidad de las Aguas. In: Domínguez E., Fernández HR. (eds.). Macroinverte‐ brados Bentónicos Sudamericanos: Sistemática y Biología. Tucumán: Fundación Mi‐

[13] Springer M. Biomonitoreo Acuático. In: Springer M., Ramírez A., Hanson P. (eds.). Macroinvertebrados de Agua Dulce de Costa Rica I. Revista de Biología Tropical

[14] Castillo LE. Pesticide impact of intensive banana production on aquatic ecosystems

[15] Castillo LE., De la Cruz E., Ruepert C. Ecotoxicology and Pesticides in Tropical Aquatic Ecosystems of Central America. Environ. Toxicol. Chem. 1997; 16: 41-51.

[16] Castillo LE., Ruepert C., Pinnock M. Acute toxicity and pesticide residues in surface waters in a pineapple growing area of Costa Rica. 10th International Symposium on

in Costa Rica. PhD thesis. Stockholm University: Stockholm; 2000.

ca. Tokio: The United Nations University; 1990.

Madison: University of Wisconsin Press; 1980.

D.F.: Plaza y Valdés, S.A de C.V.; 2000. p17-41.

Toxicity Assessment, Quebec City. 2001.

254-266.

328 30 Biodiversity in Ecosystems - Linking Structure and Function

Ganadería; 1990.

1999; 23(88): 375-387.

dad de Antioquia; 2003.

guel Lillo; 2009. P631-654.

2012; 58(4): 53-59.

2(4): 54-60.


[42] Arroyo AT. Modelos mixtos en el análisis de *loci* de caracteres cuantitativos y su apli‐ cación en mejoramiento vegetal. PhD thesis. Universidad Nacional de Córdoba: Cór‐ doba; 2008.

[30] Ramírez A., Pringle CM. Invertebrate Drift and Benthic Community Dynamics in a Lowland Neotropical Stream, Costa Rica. Hydrobiologia 1998a; 11(1-3): 19-26.

[31] Ramírez A., Pringle CM. Structure and Production of a Benthic Insect Assemblage in

[32] Ramírez A., Pringle CM. Spatial and Temporal Patterns of Invertebrate Drift in Streams Draining in a Neotropical Landscape. Freshwater Biology 2001; 46(1): 47-62.

[33] Ramírez A., Paaby P., Pringle M., Agüero G. Effect of Habitat Type on Benthic Mac‐ roinvertebrates in Two Lowland Tropical Streams, Costa Rica. Rev. Biol. Trop. 1998;

[34] Rosemond AD., Pringle CM., Ramírez A. Macroconsumer Effects on Insect Detriti‐ vores and Detritus Processing in a Tropical Stream. Freshwater Biology 1998; 39(3):

[35] Mitsch WJ., Tejada J., Nahlik A., Kohlmann B., Bernal B. Hernández C. Tropical Wet‐ lands for Climate Change Research, Water Quality Management and Conservation Education on a University Campus in Costa Rica. Ecological Engineering 2008; 34(4):

[36] Vásquez Morera A. Soils. In: Janzen DH. (ed.). Costa Rican Natural History. Chicago:

[37] Stein H., Springer M., Kohlmann B. Comparison of Two Sampling Methods for Bio‐ monitoring Using Aquatic Macroinvertebrates in the Dos Novillos River, Costa Rica. In: Kohlmann B., and Mitsch MJ. (eds.). Ecological Engineering 2008; 34(4): 267-275.

[38] MINAE (Ministerio de Ambiente y Energía, CR). Reglamento para la Evaluación y Clasificación de la Calidad de Cuerpos de Agua Superficiales. Decreto Ejecutivo Nº 33903-MINAE-S. Alcance No. 8 a La Gaceta No. 78, lunes 17 de septiembre del 2007. San José, CR. Available from: http://historico.gaceta.go.cr/pub/2007/09/17/

[39] Liess M., Brown C., Dohmen P., Duquesne S., Hart A., Heimbach F., Krueger J., Laga‐ dic L., Maund S., Reinert W., Streloke M., Tarazona JV. (eds.). Effects of pesticides in

[40] Ramírez A., Gutiérrez-Fonseca PE. Functional Feeding Groups of Aquatic Insect Families in Latin America: A Critical Analysis and Review of Existing Literature.

[41] Akaike H. Information theory and an extension of the maximum likelihood principle. Tsahkadsor, Armenian SSR. 2nd International Symposium on Information Theory;

the field. Le Croisic: EU and SETAC Europe Workshop; 2003.

Rev. Biol. Trop. 2014; 62(Suppl. 2): 155-167.

a Neotropical Stream. J. N. Am. Benthol. Soc. 1998b; 17(4): 443-463.

46(Supl. 6): 201-213.

330 32 Biodiversity in Ecosystems - Linking Structure and Function

The University of Chicago Press; 1983.

COMP\_17\_09\_2007.html

515-523.

276-288.

1973.


[67] Hilsenhoff WL. Rapid Field Assessment of Organic Pollution with a Family-Level Bi‐ otic Index. Journal of the North American Benthological Society 1988; 7: 65-68.

[56] Tomanova S., Tedesco PA. Tamaño Corporal, Tolerancia Biológica y Potencial de Bi‐ oindicación de la Calidad del Agua de *Anacroneuria* spp. (Plecoptera: Perlidae) en

[57] Thorp JH., Covich AP. Ecology and Classification of North American Freshwater In‐

[58] O´Callaghan P., Kelly-Quinn M. Performance of Selected Macroinvertebrate-Based Biotic Indices for Rivers Draining the Merendón Mountains Region of Honduras.

[59] Montoya Moreno Y., Acosta García Y., Zuluaga Zuluaga E., García A. Evaluación de la Biodiversidad de Insectos Acuáticos y de Calidad Fisicoquímica y Biológica del Río Negro (Antioquia-Colombia). Revista Universidad Católica de Oriente 2007; 23:

[60] Sánchez Argüello R., Cornejo A., Boyero L., Santos Murgas A. Evaluación de la Cali‐ dad del Agua en la Cuenca del Río Capira, Panamá. Tecnociencia 2010; 12(2): 57-70.

[61] Miserendino ML., Pizzolón LA. Rapid Assessment of River Water Quality Using Macroinvertebrates: A Family Level Biotic Index for the Patagonic Andean Zone. Ac‐

[62] Mafla MH. Guía para Evaluaciones Ecológicas Rápidas con Indicadores Biológicos en Ríos de Tamaño Medio Talamanca – Costa Rica. Macroinvertebrados (BMWP – CR – Biological Monitoring Working Party) y Hábitat (SVAP – Stream Visual Assessment Protocol). Turrialba: Centro Agronómico Tropical de Investigación y Enseñanza

[63] Macchi PA. Calidad del agua en ecosistemas fluviales utilizando macroinvertebrados bentónicos. Cuenca del arroyo Pocahullo, San Martín de los Andes. BSc thesis. Uni‐

[64] Fenoglio S., Badino G., Bona F. Benthic Macroinvertebrate Communities as Indicators of River Environment Quality: An Experience in Nicaragua. Rev. Biol. Trop. 2002;

[65] Ghetti RE. Manuale di Applicazione dell'Indice Biotico Esteso (I.B.E.). Trento: Pro‐

[66] Sermeño Chicas JM., Serrano Cervantes L., Springer M., Paniagua Cienfuegos MR., Pérez D., Rivas Flores AW., Menjívar Rosa, RA., Bonilla de Torres BL., Carranza Es‐ trada FA., Flores Tensos JM., Gonzáles C. de los A., Gutiérrez Fonseca PE., Hernán‐ dez Martínez MA., Monterrosa Urías AJ., Arias de Linares AY. Determinación de la Calidad Ambiental de las Aguas de los Ríos de El Salvador, Utilizando Invertebrados Acuáticos: Índice Biológico a Nivel de Familias de Invertebrados Acuáticos en El Sal‐

vador (IBF-SC-2010). San Salvador: Editorial Universitaria (UES); 2010.

América del Sur. Revista de Biología Tropical 2007; 55(1): 67-81.

Cuadernos de Investigación UNED 2013; 5(1): 45-54.

ta Limnologica Brasiliensia 1999; 11: 137-148.

versidad Nacional del Comuahue, Neuquén; 2007.

vincia Autonoma di Trento Press; 1997

vertebrates. Academic Press; 2010.

332 34 Biodiversity in Ecosystems - Linking Structure and Function

71-84.

(CATIE); 2005.

50(3-4): 1125-1131.


## **Ecohidrology and Nutrient Fluxes in Forest Ecosystems of Southern Chile**

Carlos E Oyarzún and Pedro Hervé-Fernandez

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/59016

## **1. Introduction**

Nitrogen (N) cycling in terrestrial ecosystems is a global environmental concern. The N cycle is a complex interplay where biotic and abiotic processes interact to transform and transfer N in an ecosystem. In general, one can simplify by classifying terrestrial N cycles all over the world in two groups: 'tight' N cycles and 'open' N cycles. The 'tight' N cycle is characterized by its high efficiency in producing bioavailable N and retaining it in the plant-soil system. The 'open' N cycle, on the other hand, is then considered to be less efficient, showing significant loss of N towards aquatic ecosystems and the atmosphere. The latter losses might lead to adverse effects on stream water and air quality, contributing as such to 'global change' [1].

The movement of nutrients between ecosystems is called geochemical cycling or external cycling. Two important input processes to forests are atmospheric deposition and mineral weathering [2]. The atmospheric input to forests consists of dry, and wet deposition. Aerosol and gases can by deposited directly from the air to plant and soil surfaces during rainless periods by dry deposition. Wet deposition is defined as the input of atmospheric compounds to the earth´s surface by rain, hail, snow and/or occult deposition that occurs via fogs and clouds, which can be important in mountainous regions [3]. During rain events, dry deposition is washed off from plant parts and, together with wet deposition, reaches the forest floor as throughfall and stem flow. A second input process is the weathering of soil minerals as a result of chemical dissolution. In combination with atmospheric deposition, mineral weathering is the only long-term source of base cations for terrestrial ecosystems [2].

The temperate climate region of southern Chile still reflects undisturbed, pre-industrial environmental conditions [4]. This is in strong contrast with land use, which has been altered significantly over the last decades and centuries. Only fragments of the original forest vegetation remain unaltered, and are located in the Coastal and Andes mountain ranges (CMR

<sup>© 2015</sup> The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and eproduction in any medium, provided the original work is properly cited.

and AMR, respectively). Exotic tree plantations and agricultural areas dominate the central valley of southern Chile [5]. These characteristics make this region an ideal study area to investigate human impacts on biogeochemical nutrient cycling. Temperate forests in Chile are not yet affected by elevated N deposition, as is the case for forests in Europe or northeastern North America [6]. However, anthropogenic activities such as transport, industry and agriculture have been increasing in central and southern Chile. These activities can substan‐ tially alter the atmospheric N load and enhance N input on forest ecosystems in Chile [5].

Several biogeochemical studies have been carried out most in humid temperate forest ecosystems between 40° and 43° S in southern Chile [i.e. 7; 8; 9]. The annual mean temperature is 5 to 12° C and precipitation ranges from 2000 to 7000 mm in the AMR [3]. Data from [5] reported that mean annual N composition of the rainwater in the CMR and AMR ranges (41°-43° S), varied between < 30 – 43 NO3 - -N µg L-1 and 9.8 – 26.2 NO3 - -N µg L-1. Similarly, NH4 + -N concentrations were < 50 NH4 + -N µg L-1 and between 39.5 – 45.4 NH4 + -N µg L-1 for CMR and AMR, respectively. Forests in the CMR, are located immediately near the ocean and are unique in this sense that external input of major elements are almost exclusively due to marine aerosols. Since trees canopy act as efficient filters, forests can capture large amounts of atmospheric deposition, especially occult deposition (i.e: fog and cloud). Normally, mountain forest ecosystems are very efficient in trapping nutrients, especially N and cations from clouds and fogs [10; 11; 4].

Stream nutrient loads are heavily dependent on catchment vegetation. Alteration of canopies and the soil under it, have a significant impact on nitrogen (NO3 - -N; NH4 + -N; DON and TDN) and phosphorus (PO4 3+-P and TDP) reaching the stream. Human disturbances have a direct impact on biological communities and may lead to land degradation, causing a change in ecosystem services and livelihood support. Temperate rain forest ecosystems of southern Chile have efficient mechanisms of retention for essential nutrients, especially NH4 + and NO3 - [7, 3). [6] described that the dominant form of N leaching was dissolved organic nitrogen (DON) for unpolluted forests of southern Chile. Other studies in the area had reported that conversion from native forests to exotic fast-growing plantations is likely to decrease N retention on catchments [12].

### **1.1. Native temperate rainforests of southern Chile**

Native temperate rainforests of southern Chile represent an important global reserve of temperate forest with an extraordinary genetic, phytogeographic and ecological significance [13] with a worldwide high conservation priority [14]. These forests cover an area of 13.5 million ha. and are isolated by physical and climatic barriers, resulting in high endemism in plants and animals: 28 of 82 genera of woody plants (34%) are endemic to the region, along with 50% of vines, 53% of hemiparasites and 45% of vertebrates [15]. Some taxa are derived from ancient elements in southern Gondwana. Some relict tree species of conifers have the longest recorded lifespan, reaching an age of up to 3,600 years, constituting an excellent historical document for studies in reconstruction of climatic variability [16]. Most of the Valdivian eco-region is also considered as part of the world's 25 hotspots for biodiversity conservation and some of its forest types are included among the last frontier forests in the planet. These forests support fundamental ecological functions, which provide a range of ecosystem services and goods such as conservation of biological diversity, maintenance of soil fertility, and timber and non-timber products [17]. Also they contribute to maintain fresh water supply, which in turn supports the availability of drinkable water for cities [18].

Native forests in the Valdivian eco-region (36° S through 48° S) have suffered anthropical disturbances due to inadequate logging practices, and to agricultural land or exotic fast growing plantations conversion. Rapid conversion to forest plantations between 1975 and 2000 resulted in deforestation rates of 4.5% per year within an area of 578,000 ha in the Maule region (38° S), facilitated through afforestation incentives [19]. Another important cause of defores‐ tation has been human-set fires, with an annual average of 13,000 ha burned in the period 1995– 2005 and a high interannual variability associated to rainfall variation [20]. Anthropogenic land cover change in the central depression of southern Chile (40°-42° S) is the most evident process of deforestation and agricultural expansion. A large fraction of the *Nothofagus* forests in that region has been cleared for agriculture during the last century [21]. Patches of secondgrowth forest cover vast areas of the regional landscape, leaving only scattered stands as a result from intensified agriculture activity. Direct effects of past land use may occur via longterm (> 50 yr) physical alteration of the rhizosphere caused by historic practices. Soil compac‐ tion is an enduring consequence of cultivation, grazing, and logging that can cause increased bulk density and reduced pore space [1]. These changes may affect the abundance of aerobic and anaerobic microorganisms and subsequently reduce the cycling of several elements, including N.

### **1.2. Eucalyptus plantation forests**

and AMR, respectively). Exotic tree plantations and agricultural areas dominate the central valley of southern Chile [5]. These characteristics make this region an ideal study area to investigate human impacts on biogeochemical nutrient cycling. Temperate forests in Chile are not yet affected by elevated N deposition, as is the case for forests in Europe or northeastern North America [6]. However, anthropogenic activities such as transport, industry and agriculture have been increasing in central and southern Chile. These activities can substan‐ tially alter the atmospheric N load and enhance N input on forest ecosystems in Chile [5].

Several biogeochemical studies have been carried out most in humid temperate forest ecosystems between 40° and 43° S in southern Chile [i.e. 7; 8; 9]. The annual mean temperature is 5 to 12° C and precipitation ranges from 2000 to 7000 mm in the AMR [3]. Data from [5] reported that mean annual N composition of the rainwater in the CMR and AMR ranges

CMR and AMR, respectively. Forests in the CMR, are located immediately near the ocean and are unique in this sense that external input of major elements are almost exclusively due to marine aerosols. Since trees canopy act as efficient filters, forests can capture large amounts of atmospheric deposition, especially occult deposition (i.e: fog and cloud). Normally, mountain forest ecosystems are very efficient in trapping nutrients, especially N and cations from clouds

Stream nutrient loads are heavily dependent on catchment vegetation. Alteration of canopies

impact on biological communities and may lead to land degradation, causing a change in ecosystem services and livelihood support. Temperate rain forest ecosystems of southern Chile

[6] described that the dominant form of N leaching was dissolved organic nitrogen (DON) for unpolluted forests of southern Chile. Other studies in the area had reported that conversion from native forests to exotic fast-growing plantations is likely to decrease N retention on

Native temperate rainforests of southern Chile represent an important global reserve of temperate forest with an extraordinary genetic, phytogeographic and ecological significance [13] with a worldwide high conservation priority [14]. These forests cover an area of 13.5 million ha. and are isolated by physical and climatic barriers, resulting in high endemism in plants and animals: 28 of 82 genera of woody plants (34%) are endemic to the region, along with 50% of vines, 53% of hemiparasites and 45% of vertebrates [15]. Some taxa are derived from ancient elements in southern Gondwana. Some relict tree species of conifers have the longest recorded lifespan, reaching an age of up to 3,600 years, constituting an excellent historical document for studies in reconstruction of climatic variability [16]. Most of the Valdivian eco-region is also considered as part of the world's 25 hotspots for biodiversity conservation and some of its forest types are included among the last frontier forests in the


3+-P and TDP) reaching the stream. Human disturbances have a direct






and NO3 - [7, 3).

+


+


+

have efficient mechanisms of retention for essential nutrients, especially NH4

and the soil under it, have a significant impact on nitrogen (NO3

**1.1. Native temperate rainforests of southern Chile**

(41°-43° S), varied between < 30 – 43 NO3

3362 Biodiversity in Ecosystems - Linking Structure and Function


NH4 +

and fogs [10; 11; 4].

and phosphorus (PO4

catchments [12].

In south-central Chile (35-40° S), the native vegetation has been converted to agricultural uses, primarily plantation forestry, which has resulted in a landscape dominated by industrial forestry plantations. The amount of land in the region classified as plantation forestry has increased by 55 % between 1998 and 2008 (116–179 thousands ha; [22]. As in other parts of Chile, over 20,000 ha of those new plantations have replaced native forests in the region [19, 23], mainly located in the CMR. The growth of exotic species in non-native environments has uncertain ecohydrological consequences [24]. Therefore, there is much concern about their water consumption. Several authors have concluded that the consequences of exotic fast growing plantations are: (i) the decrease of discharge due to higher evapotranspiration [25, 26]; and (ii) changes in the soil hydrological properties, such as infiltration rates [27] and soil hydrophobicity [28].

## **2. Objectives**

In small headwater catchments located at the Costal mountain range (CMR), in southern Chile (40° S), concentrations and fluxes of NO3 - -N, NH4 + -N, DON, TDN, TDP and base cations (Ca2+, Mg2+, Na+ and K+ ) in bulk precipitation, throughfall and catchment discharge water were measured. The main objective of this study was to compare how hydrological variability affects catchment nutrient load responses with different land cover of native forests and exotic plantation of *Eucalyptus spp.*, in order to evaluate possible effects of land use

## **3. Material and methods**

### **3.1. Description of the study sites**

We selected five catchments with different land cover: (a) one with old-growth native ever‐ green rainforest (ONE), (b) one with native deciduous *Nothofagus obliqua* forest (ND), (c) one with secondary native evergreen forest (NE), (d) one covered with exotic fast growing *Eucalyptus nitens* (FEP) and (e) one with fast-growing exotic cover of *Eucalyptus globulus* (EG), located at CMR (40°S), near the city of Valdivia, Chile. All five catchments are located inland from the Pacific coast. The ONE catchment has an area of 2.8 ha at 336 m a.s.l. and is 20 km from the coast. The ND catchment has an area of 10.1 ha at 71-125 m a.s.l., and is 23.0 km from the coast. The NE catchment has an area of 3.1 ha at 227-275 m a.s.l. and is 2.0 km from the coast. The FEP catchment has an area of 54.8 ha and is 18 km from the coast, and the EG catchment has an area of 5.6 ha at 250-297 m a.s.l., and is 2.6 km from the coast.

### **3.2. Forest cover**

In the catchment covered by old-growth native evergreen rainforest (ONE) the main canopy species are *Eucryphia cordifolia* Cav., *Aextoxicon punctatum* Ruiz et Pav. and *Laureliopsis philippiana* (Looser) Schodde. This last shows the highest density (718 tree ha-1 ) and basal area (37.2 m2 ha-1) (Figure 1). The understorey is dominated by *Amormyrtus luma*, *Amomyrtus meli*, *Drimys winteri* and *Myrceugenia planipes*. The attributes of the old-growth native rainforests in the study area includes: increase in the proportion of successional species, the promotion of better growth rates to reach large diameters, the development of a rich understory and new regeneration cohorts, the increase of vertical structure, the development of increased wildlife habitat, and the presence of dead wood in the system (snags, and coarse woody debris) [29].

The main canopy species in the mixed ND catchment is the deciduous species *Nothofagus obliqua* (Mirb.) Oerst. reaching heights of 35 m, which covers 63.3 % of the catchment. Also, 13.8 and 7.9 percent is covered by native secondary forests of *Gevuina avellana* and *Astrocedrus chilensis* planted in 1983 and 1982, respectively, and 15.0 percent is covered by the fast-growing *Eucalyptus sp.* plantation. Understorey trees include *Luma apiculata*, *Podocarpus salignus*, *Aextoxicon punctatum*, *Amomyrtus meli*, *Gevuina avellana* and the exotic tree *Acacia melanoxylon*. Shrubs that reach heights over 3 m are mainly *Chusquea quila* Kunth with a 95% canopy cover.

In the NE catchment, the vegetation cover is characterized as a second growth native evergreen forest, dominated by *Myrtaceae* spp., *Amomyrus luma* (Mol.) Legr. et. Kaus (29%), *Amomyrtus meli* (Phil.) Legr. et. Kaus (25%), *Laureliopsis phillipiana* (Mol.) Mol. (14%), *Myrceugenia pla‐ nipes* (Hook. et Arn.) Berg. (13%), *Dyasaphillum diacanthoides* (Less.) Cabrera (7%), *Gevuina avellana* (Molina) Molina (6%), *Lomatia ferrugina* (Cav.) R. Br., *Persea lingue* (Ruiz et Pav.) Nees ex Koop. and *Myrceugenia exucca* (DC.) Berg. (2% each) and *Aextoxicon punctatum* (1%). This catchment is also used as a source of wood by local residents and as an occasional grazing ground for animals during the winter.

catchment nutrient load responses with different land cover of native forests and exotic

We selected five catchments with different land cover: (a) one with old-growth native ever‐ green rainforest (ONE), (b) one with native deciduous *Nothofagus obliqua* forest (ND), (c) one with secondary native evergreen forest (NE), (d) one covered with exotic fast growing *Eucalyptus nitens* (FEP) and (e) one with fast-growing exotic cover of *Eucalyptus globulus* (EG), located at CMR (40°S), near the city of Valdivia, Chile. All five catchments are located inland from the Pacific coast. The ONE catchment has an area of 2.8 ha at 336 m a.s.l. and is 20 km from the coast. The ND catchment has an area of 10.1 ha at 71-125 m a.s.l., and is 23.0 km from the coast. The NE catchment has an area of 3.1 ha at 227-275 m a.s.l. and is 2.0 km from the coast. The FEP catchment has an area of 54.8 ha and is 18 km from the coast, and the EG

In the catchment covered by old-growth native evergreen rainforest (ONE) the main canopy species are *Eucryphia cordifolia* Cav., *Aextoxicon punctatum* Ruiz et Pav. and *Laureliopsis*

*Drimys winteri* and *Myrceugenia planipes*. The attributes of the old-growth native rainforests in the study area includes: increase in the proportion of successional species, the promotion of better growth rates to reach large diameters, the development of a rich understory and new regeneration cohorts, the increase of vertical structure, the development of increased wildlife habitat, and the presence of dead wood in the system (snags, and coarse woody debris) [29].

The main canopy species in the mixed ND catchment is the deciduous species *Nothofagus obliqua* (Mirb.) Oerst. reaching heights of 35 m, which covers 63.3 % of the catchment. Also, 13.8 and 7.9 percent is covered by native secondary forests of *Gevuina avellana* and *Astrocedrus chilensis* planted in 1983 and 1982, respectively, and 15.0 percent is covered by the fast-growing *Eucalyptus sp.* plantation. Understorey trees include *Luma apiculata*, *Podocarpus salignus*, *Aextoxicon punctatum*, *Amomyrtus meli*, *Gevuina avellana* and the exotic tree *Acacia melanoxylon*. Shrubs that reach heights over 3 m are mainly *Chusquea quila* Kunth with a 95% canopy cover.

In the NE catchment, the vegetation cover is characterized as a second growth native evergreen forest, dominated by *Myrtaceae* spp., *Amomyrus luma* (Mol.) Legr. et. Kaus (29%), *Amomyrtus meli* (Phil.) Legr. et. Kaus (25%), *Laureliopsis phillipiana* (Mol.) Mol. (14%), *Myrceugenia pla‐ nipes* (Hook. et Arn.) Berg. (13%), *Dyasaphillum diacanthoides* (Less.) Cabrera (7%), *Gevuina avellana* (Molina) Molina (6%), *Lomatia ferrugina* (Cav.) R. Br., *Persea lingue* (Ruiz et Pav.) Nees ex Koop. and *Myrceugenia exucca* (DC.) Berg. (2% each) and *Aextoxicon punctatum* (1%). This

ha-1) (Figure 1). The understorey is dominated by *Amormyrtus luma*, *Amomyrtus meli*,

) and basal area

plantation of *Eucalyptus spp.*, in order to evaluate possible effects of land use

catchment has an area of 5.6 ha at 250-297 m a.s.l., and is 2.6 km from the coast.

*philippiana* (Looser) Schodde. This last shows the highest density (718 tree ha-1

**3. Material and methods**

**3.2. Forest cover**

(37.2 m2

**3.1. Description of the study sites**

3384 Biodiversity in Ecosystems - Linking Structure and Function

The FEP catchment is covered with *Eucalyptus nitens* of 4 and 14 yr-old. However, this catch‐ ment has had already 5 *E. nitens* rotations; density of 2911 tree ha-1 and the basal area is 131.9 m2 ha-1. In FEP, the highest density was observed in the diameter 20-25 and 25-30 cm (Figure 2). The totaldensity ranges between 2911-2733 treeha-1 in the sites.Basal area ranges between 131.9 and 144.4 m2 ha-1, and the mean height of the trees was 25.4 m. The riparian vegetation of the catchment with *Eucalyptus nitens* plantation has a large proportion of small trees and shrubs with a diameter distribution between 5-10 cm (Figure 3). The main tree species is *Luma apiculata* with 2180 tree ha-1 and the shrub *Aristotelia chilensis* with 815 tree ha-1 (Figure 3).

In EG catchment, the vegetation cover is composed of 80% exotic plantation of *Eucalyptus globulus* and 20% native evergreen remnant as a buffer zone. This is composed of *Berberis darwini* (Hooker) and *Ovidia pillopillo* (Gray) Hohen ex Meissn. (both with 29% cover), *Eucriphya cordifolia* Cav. (25.8%), *Lomatia ferruguinea* (Cav.) R. Br. (9.7%), *Dasyphyllum diacanthoides*(Less.) Cabrera and *Raphitamnus spinosus* (Juss.) Mold. (both with 3.2%). Originally this catchment was a native evergreen forest. However it was cleared (35 years ago) with fire to open areas for grazing animals, and in some areas, for the extraction of wood. Recently (9 years ago) the grassland was replaced by exotic trees (*Eucalyptus globulus*). Local residents use the forest as a source of wood and also allow animals to graze on the grass as well as on tree shoots.

**Figure 1.** Diameter distributions by species (AL=*Amomyrtus luma*, AM=*Amomyrtus meli*, AP=*Aextoxicon punctatum*, EC=*Eucryphia cordifolia,* GA=*Gevuina avellana,* LP=*Laureliopsis philippiana*, OE=other species) in the catchment with na‐ tive old-growth evergreen rainforests.

**Figure 2.** Diametric classes of species found on FEP and EG catchments.

**Figure 3.** Diametric classes and density of trees for ONE and ND. Note that since ONE had the oldest trees, the last two classes comprise trees within 45 to 100, and 100 to 190 cm diameter at breast high.

#### **3.3. Soils and climate**

Climate in the area of study, is rainy temperate. In the meteorological station Isla Teja (25 m a.s.l.), 10 to 20 km from the study sites, the mean annual temperature is 12.0 °C (January mean is 17 °C and July mean is 7.6 °C) and the mean annual precipitation is 2,280 mm. Rainfall is concentrated during winter (May–August, 62 %) and decreases strongly in the summer (January–March, 9 %). Soils in the study area are red clayish derivatives from ancient volcanic ashes, deposited over a metamorphic geological substratum, dominated by micaceous schist and quartz lenses. The soils are shallow (< 1.0 m depth) in EG and NE catchments, and predominantly deep (> 1.0 m) in ND catchment. Soils in the EG catchment are characterized by poor infiltration rates, and in the NE and ND catchments by high infiltration rates [27].

Soils at ONE and FEP catchments have approximately the same texture in the bottom of the 1 meter depth soil profile, however the top layers (0 to 15; and 15 to 30) have consistently 10% more clay, and 1% less sand in FEP compared to ONE soil profiles. In the FEP catchment, clay content ranges between 37.2 – 45.1 %, organic matter content ranges between 1.8 – 17.1%, inorganic-N (NO3 - -N and NH4 + -N) ranges between 9.8 – 21.0 mg kg-1, Ca2+between 0.19 – 0.23 cmol kg-1 and Mg2+ranges between 0.09 – 0.16 cmol kg-1. While, ONE soil clay content ranges between 31.1 – 37.3 % and organic matter content ranges between 5.9 – 17.8 %, inorganic-N ranges between 11.2 – 57.4 mg kg-1, Ca2+ranges between 0.23 – 1.32 cmol kg-1 and Mg2+ranges between 0.10 – 0.71 cmol kg-1.

## **4. Methods**

FEP

Density (ind ha

Density (ind ha



< 5

ONE

< 5

**3.3. Soils and climate**

0

400

800

1200

5-10

10-15

15-20

20-25

25-30

30-35

35-40

two classes comprise trees within 45 to 100, and 100 to 190 cm diameter at breast high.

40-45

45-100

*A. meli A. punctatum E. cordifolia G. avellana L. philippiana M. planipes R. spinosus Other spp.* 

100-190

**Figure 3.** Diametric classes and density of trees for ONE and ND. Note that since ONE had the oldest trees, the last

Climate in the area of study, is rainy temperate. In the meteorological station Isla Teja (25 m a.s.l.), 10 to 20 km from the study sites, the mean annual temperature is 12.0 °C (January mean

DBH (cm)

0

400

800

1200

5-10

10-15

3406 Biodiversity in Ecosystems - Linking Structure and Function

15-20

20-25

**Figure 2.** Diametric classes of species found on FEP and EG catchments.

25-30

30-35

35-40

*E.nitens A. chilensis E. globulus L. apiculata O. pillopillo P. radiata Other spp.* 

40-45

DBH (cm)

ND

< 5

5-10

10-15

15-20

20-25

25-30

30-35

35-40

40-45

< 5

5-10

10-15

15-20

20-25

25-30

30-35

35-40

*E. globulus E. cordifolia O. pillopillo L. ferruginea B. darwini D. diacanthoides R. spinosus* 

40-45

*G. avellana L. ferruginea P. lingue L. apiculata A. meli A. punctatum M. planipes M. exxuca L. philippiana D. diacanthoides* 

EG

Bulk precipitation was sampled using four plastic rain collectors attached to a 2.5-liter bottle. Bulk precipitation collectors (surface area 200 cm2 ,) were installed in open areas (no trees were within 20 m of the sampling point), located between a distance of 100 – 500 m. Throughfall water was collected, using 2-4 collectors (surface area 254 cm2 ) were installed inside each type forest. All collectors were installed 1.2 m above the forest floor and installed inside opaque tubes in order to avoid light penetration that could promote algae growth. Throughfall collectors had a thin mesh at the beginning of the neck of the funnel, in order to prevent insects and leaves entering the collection bottles, and designed with a plastic ring in order to exclude bird droppings [30]. Soil water was sampled at two different depths (0.3, 0.6 m) with lowtension porous-cup lysimeters (max 60 kPa of tension was applied) (Soil Moisture equipment corp.).

Discharge from each catchment was constantly measured by a pressure transducer paired with a baro diver (Schlumberger Water Services). Water samples were taken directly from the streams with an ISCO-6712 automatic sampler in each catchment. Stream samples were composed by two 250 mL aliquots taken each 30 minutes (1 h compound sample per bottle). Samples were filtered through a borosilicate glass filter (Whatman) of 0.45 µm. NO3 - -N (NO3 - - N+NO2 - -N) was determined by the cadmium reduction method, where NO2 - -N was always below detection limits. NH4 + -N was determined with the phenate method (blue indophenol), detection limit (DL) was < 2 µg L-1, for nitrite, nitrate and ammonia. Dissolved Inorganic Nitrogen (DIN) was calculated as follows: DIN=NO3 - -N+NO2 - -N+ NH4 + -N. Total dissolved nitrogen (TDN) was determined by the sodium hydroxide and persulfate digestion method (DL < 15 µg L-1). Organic nitrogen (DON) was calculated by subtracting (DON=TDN-DIN) concentration from TDN. Total dissolved phosphorous (TDP) was measured by the sodium hydroxide and persulfate digestion method (DL < 3 µg L-1) at LIMNOLAB (Limnology Laboratory, Universidad Austral de Chile). Ca2+and Mg2+(± 0.05 mg L-1) were analyzed by AAS, while Na+ and K+ (± 0.05 mg L-1) by AES in the Forestry Nutrition and Soil Laboratory, Univer‐ sidad Austral de Chile.

Canopy enrichment factors were calculated as the ratio between throughfall and bulk precip‐ itation from different forest covers (throughfall / bulk precipitation). Fluxes were calculated using discharge and rainfall volumes. While nutrient retention (R) was calculated as follows:

> ( ) ® Retention = Input – Output / Input Where, R > 0, + Retention R = 0, Equilibrium R < 0, - Retention

## **5. Results and discussion**

### **5.1. Throughfall enrichment factors**

Canopy enrichment factors are presented in Figure 4. ND and ONE forests showed the highest enrichment and variability, whereas the EG plantation showed the lowest. The nutrient which presentedthelowestannualenrichmentinallthroughfall sampleswasNO3 - -Nrangingfrom-0.8 for EG, through 1.5 for FEP. The highest enrichment was DON (10.3 times) for ONE and TDP (10.7 times) for ND forests. This enrichment is due to two processes: the washing off of the unquantified N input by dry deposition, on the one hand, and the N uptake from wet, dry particulate and gaseous deposition by leaves, twigs, stem surfaces, and lichens, on the other hand[31].The old-growth evergreenforests (likeONEcatchment) aremulti-stratifiedandhave an understory of high diversity, resulting in a complex and diverse structure and species composition. Also, [32] reported that DIN and DON concentrations were higher in through‐ fall than in bulk precipitation, particularly for nitrate, in a native *Nothofagus obliqua* forest and a *Pinus radiata* plantation, located near of the study sites. [8] observed 3.7 times throughfall enrichment for NO3 - -N, in an evergreen *Nothofagus betuloides* forest (9.8 µg L-1 and 36.5 µg L-1 for bulk precipitation and throughfall, respectively) and a 1.7 throughfall enrichment under a deciduous *Nothofagus pumilio* forest (26.2 µg L-1 and 43.5 µg L-1 for bulk precipitation and throughfall,respectively) at cordillera de los Andes (40° S, 1120 m a.s.l.). However, NH4 + -N was retained by canopies. Data from forested sites in the USA and Europe [33] showed that net canopy exchange of N (throughfall plus stemflow minus bulk precipitation) was negative for NH4+and NO3-at all sites, indicating that canopies were clearly sinks for inorganic N.

### **5.2. Annual nutrient fluxes**

TDN annual retention and net annual fluxes (in kg N ha-1 yr-1) was 0.58 (1.43); 0.90 (9.31) and-4.79 (-7.14) for NE, ND and EG forests, respectively. TDP annual retention and net annual fluxes (in kg P ha-1 yr-1) were 0.70 (0.08); 0.96 (0.06) and-1.44 (0.4) for NE, ND and EG, respec‐ tively (Figure 4). Studies in watersheds in the United States [34] reported that thin or porous soils and high infiltration rates have less capacity to retain N. However, in our study, catch‐ ments with high infiltration rates, such as NE and ND showed greater N retention than soils with very low infiltration rates, such as EG. In our study, the differences in DIN retention were evident between native forests and *Eucalyptus* plantation, as also has been described previously by [12]. However, [35] observed using land cover, watershed area and precipitation as predictors for water quality (nitrate, ammonia, DON, TDP and electric conductivity) for local models explained 79.5% of the variance.

 **Figure 4.** Throughfall enrichment factors for the five catchments (left) and annual nutrient fluxes for three catchments (right). EG=*Eucalyptus globulus* plantation, NE=native secondary evergreen ND=native deciduous, ONE=native oldgrowth evergreen, FEP=*Eucalyptus nitens* plantation.

#### **5.3. Nutrient concentration in stream water**

Throughfall Enrichment Factor

concentration from TDN. Total dissolved phosphorous (TDP) was measured by the sodium hydroxide and persulfate digestion method (DL < 3 µg L-1) at LIMNOLAB (Limnology Laboratory, Universidad Austral de Chile). Ca2+and Mg2+(± 0.05 mg L-1) were analyzed by AAS,

Canopy enrichment factors were calculated as the ratio between throughfall and bulk precip‐ itation from different forest covers (throughfall / bulk precipitation). Fluxes were calculated using discharge and rainfall volumes. While nutrient retention (R) was calculated as follows:

( )

Retention = Input – Output / Input Where, R > 0, + Retention R = 0, Equilibrium R < 0, - Retention

Canopy enrichment factors are presented in Figure 4. ND and ONE forests showed the highest enrichment and variability, whereas the EG plantation showed the lowest. The nutrient which

for EG, through 1.5 for FEP. The highest enrichment was DON (10.3 times) for ONE and TDP (10.7 times) for ND forests. This enrichment is due to two processes: the washing off of the unquantified N input by dry deposition, on the one hand, and the N uptake from wet, dry particulate and gaseous deposition by leaves, twigs, stem surfaces, and lichens, on the other hand[31].The old-growth evergreenforests (likeONEcatchment) aremulti-stratifiedandhave an understory of high diversity, resulting in a complex and diverse structure and species composition. Also, [32] reported that DIN and DON concentrations were higher in through‐ fall than in bulk precipitation, particularly for nitrate, in a native *Nothofagus obliqua* forest and a *Pinus radiata* plantation, located near of the study sites. [8] observed 3.7 times throughfall

for bulk precipitation and throughfall, respectively) and a 1.7 throughfall enrichment under a deciduous *Nothofagus pumilio* forest (26.2 µg L-1 and 43.5 µg L-1 for bulk precipitation and

retained by canopies. Data from forested sites in the USA and Europe [33] showed that net canopy exchange of N (throughfall plus stemflow minus bulk precipitation) was negative for

TDN annual retention and net annual fluxes (in kg N ha-1 yr-1) was 0.58 (1.43); 0.90 (9.31) and-4.79 (-7.14) for NE, ND and EG forests, respectively. TDP annual retention and net annual

throughfall,respectively) at cordillera de los Andes (40° S, 1120 m a.s.l.). However, NH4

NH4+and NO3-at all sites, indicating that canopies were clearly sinks for inorganic N.




+ -N was

presentedthelowestannualenrichmentinallthroughfall sampleswasNO3

®

(± 0.05 mg L-1) by AES in the Forestry Nutrition and Soil Laboratory, Univer‐

while Na+

and K+

3428 Biodiversity in Ecosystems - Linking Structure and Function

**5. Results and discussion**

enrichment for NO3

**5.2. Annual nutrient fluxes**

**5.1. Throughfall enrichment factors**


sidad Austral de Chile.

Nitrogen and phosphorous concentrations in stream water are variable in forest ecosystems of southern Chile (see Table 1). In general, the highest values of TDN and TDP concentrations are in *Fitzroya cuppressoides* forest (176.5 µg N L-1) located in Coastal mountain range and in *Nothofagus pumilio* forest (67.3 µg P L-1) located in Andean mountain range. The lowest values were found in an evergreen forest (36.8 µg N L-1), located in Coastal mountain range and in *Fitzroya cuppresoides* forest (4.6 µg P L-1), and located in the Coastal mountain range. Concen‐ trations of inorganic N were smaller in the evergreen forest (33.2 µg L-1) and in *E. nitens* plantation (33.6 µg L-1) compared to organic N (94.4 and 67.0 µg L-1, respectively), in agreement with previous research in southern Chile [6; 3] demonstrating that dissolved organic nitrogen is responsible for the majority of nitrogen losses from unpolluted forest ecosystems.


**Table 1.** Mean concentrations (µg L-1) of TDN and TDP in stream water for different forest ecosystems under a lowdeposition climate, southern Chile. At the end of the table 1, is the average for each location: Andean mountain range (AMR) and Coastal mountain range (CMR).

### **5.4. Relationships between discharge and nutrient concentrations**

Nutrient exportation is related to hydrology, since water transports chemical compounds and particles. The relations of TDN and TDP with catchment discharge were positive for all nutrients except DIN, which showed a negative relation with discharge, during wet season (Figure 5). This negative relation is due to the dilution of nitrate with rainfall water which has higher concentrations of NH4 + -N.

For dry season, the fitted models showed relatively high adjusted r2 values for the *E. nitens* covered catchment for TDN and TDP (0,952 and 0,826, respectively; both with p < 0.05). However, the old growth covered catchment showed much lower values for TDN and TDP (0.317 and 0.519, respectively). Nevertheless, only TDP was significant. Dry season event DIN exportation was best fitted with a linear model. However, the fit was poor and not significant for both catchments. During wet season, the adjusted r2 values were higher for *E. nitens* covered catchment than the old growth covered catchments (Table 2). On figure 5, is clearly seen that during dry season TDN, TDP and DIN increase rapidly as discharge increases in *E.* *nitens* covered catchment (FEP). However this is not observed for the old growth covered catchment (ONE). However, during wet season TDP shows greater increase in concentrations in ONE, rather than FEP. TDN and DIN shows the same behaviour in both catchments.

**Type of forest Forest description Location TDN TDP References**

Native deciduous *N.nervosa-N. obliqua* AMR 73.3 44 Unpublished Native evergreen Evergreen forest AMR 157 18 Unpublished Native evergreen Evergreen forest AMR 67.3 37.4 Unpublished

Native conifer *Fitzroya cuppressoides* CMR 177 4.6 [9] Native evergreen Evergreen forest CMR 36.8 24.1 [12]

Native evergreen Evergreen forest CMR 127 11.1 Unpublished

Exotic monoculture *Eucalyptus nitens* CMR 100 11 Unpublished

**Table 1.** Mean concentrations (µg L-1) of TDN and TDP in stream water for different forest ecosystems under a lowdeposition climate, southern Chile. At the end of the table 1, is the average for each location: Andean mountain range

Nutrient exportation is related to hydrology, since water transports chemical compounds and particles. The relations of TDN and TDP with catchment discharge were positive for all nutrients except DIN, which showed a negative relation with discharge, during wet season (Figure 5). This negative relation is due to the dilution of nitrate with rainfall water which has

For dry season, the fitted models showed relatively high adjusted r2 values for the *E. nitens* covered catchment for TDN and TDP (0,952 and 0,826, respectively; both with p < 0.05). However, the old growth covered catchment showed much lower values for TDN and TDP (0.317 and 0.519, respectively). Nevertheless, only TDP was significant. Dry season event DIN exportation was best fitted with a linear model. However, the fit was poor and not significant for both catchments. During wet season, the adjusted r2 values were higher for *E. nitens* covered catchment than the old growth covered catchments (Table 2). On figure 5, is clearly seen that during dry season TDN, TDP and DIN increase rapidly as discharge increases in *E.*

AMR average 85.6 30.1 CMR average 115 16.2

Native deciduous *Nothofagus dombeyi* CMR 153 nd [32] Exotic monoculture *Eucalyptus spp.* CMR 94.8 30.1 [12]

AMR 109 4.9 Unpublished

Native evergreen *S. conspicua - L.*

344 10 Biodiversity in Ecosystems - Linking Structure and Function

(AMR) and Coastal mountain range (CMR).

higher concentrations of NH4

*philippiana*

**5.4. Relationships between discharge and nutrient concentrations**

+ -N.

Native deciduous *Nothofagus pumilio* AMR nd 67.3 [8] Native deciduous *Nothofagus betuloides* AMR nd 9.2 [8] Native deciduous *Nothofagus betuloides* AMR 62 nd [3]

**Figure 5.** Total dissolved nitrogen (TDN), Total dissolved phosphorus (TDP) and Dissolved inorganic nitrogen (DIN) concentrations during one dry and wet season events (for the period March – November 2013), for the catchments cov‐ ered with old growth native evergreen (ONE, in dark red circles) and catchment covered with *Eucalyptus nitens* (FEP, inverted orange triangles).


**Table 2.** Adjusted r2 values after fitting linear (L, f=y0+a x); single parameter exponential growth (1EG, f=e(a x)); 2 parameter exponential growth (2EG, f=a e(b x)) and 3 parameter exponential growth (3EG, f=y0+a e(b\*x)) models.


**Table 3.** Ca2+and Mg2+vs. discharge during events for each catchment. Adjusted r2 values after fitting linear (L, f=y0+a x) and exponential decay (ED, f=a e(-b x)) models.

**Figure 6.** Ca2+and Mg2+concentrations vs discharge for the wet season event. Dark red dots and continuous line stands for old growth evergreen covered catchment (ONE), while inverted orange triangles and segmented line stand for *Eu‐ calyptus nitens* covered catchment (FEP).

Typically, products of mineral weathering (e.g. Ca2+and Mg2+) decline in concentration when the discharge increases caused by rainfall (stream water dilutes). This was observed during wet season event, and only in FEP, for both cations. ONE showed an increase in concentration for Ca2+and a slightly reduced concentration for Mg2+.

We observed negative correlations between stream discharge and base cations concentrations (Figure 6). Typically, products of mineral weathering (e.g. Ca2+and Mg2+) decline in concen‐ tration when the discharge increases caused by rainfall (stream water dilutes). [36] reported inverse relationship between stream discharge and concentrations of Ca2+and Mg2+. However, [37] reported that during storms, both positive and negative relationships were observed between stream discharge and Ca2+and Mg2+concentrations and in some storms an initial increase in concentration was followed by dilution. On the other hand, [38] reported in an undisturbed old-growth Chilean forest that Ca2+concentration demonstrated dilution when stream discharge increase and enhanced hydrological access occurred only for H+ . According to [39], mica schists, present in the geological substrate at the coastal mountain range, are rich in micas and minerals and contain high levels or iron and magnesium. Hence, concentration levels of magnesium in stream water probably are influenced by the geological substrate. However, the dilution and increase in concentration (on FEP and ONE, respectively) is mostly due to the dilution of stream water discharge with throughfall.

## **6. Conclusions**

**Dry season event Wet season event**

values after fitting linear (L, f=y0+a x); single parameter exponential growth (1EG, f=e(a x)); 2

**Dry season event Wet season event**

Mg+2

values after fitting linear (L, f=y0+a

0 15 30 45

Catchment TDN TDP DIN TDN TDP DIN

parameter exponential growth (2EG, f=a e(b x)) and 3 parameter exponential growth (3EG, f=y0+a e(b\*x)) models.

Catchment Ca2+ Mg2+ Ca2+ Mg2+

**Table 3.** Ca2+and Mg2+vs. discharge during events for each catchment. Adjusted r2

0 15 30 45

for Ca2+and a slightly reduced concentration for Mg2+.

*calyptus nitens* covered catchment (FEP).

x) and exponential decay (ED, f=a e(-b x)) models.

346 12 Biodiversity in Ecosystems - Linking Structure and Function

ONE nd nd 0,554 (L) 0,184 (L) FEP nd nd 0,026 (L) 0,857 (ED)

Q (L s-1)

**Figure 6.** Ca2+and Mg2+concentrations vs discharge for the wet season event. Dark red dots and continuous line stands for old growth evergreen covered catchment (ONE), while inverted orange triangles and segmented line stand for *Eu‐*

Typically, products of mineral weathering (e.g. Ca2+and Mg2+) decline in concentration when the discharge increases caused by rainfall (stream water dilutes). This was observed during wet season event, and only in FEP, for both cations. ONE showed an increase in concentration

We observed negative correlations between stream discharge and base cations concentrations (Figure 6). Typically, products of mineral weathering (e.g. Ca2+and Mg2+) decline in concen‐

**Table 2.** Adjusted r2

Ca+2

mg L-1

0,0

0,5

1,0

ONE 0,317 (L) 0,519 (3EG) 0,170 (L) 0,331 (L) 0,331 (L) 0,05 (L) FEP 0,952 (1EG) 0,826 (2EG) 0,04 (L) 0,728 (L) 0,765 (2EG) 0,388 (L)

> We conclude that the mixed-deciduous (ND) and old-growth evergreen (ONE) forests show the highest canopy enrichment for throughfall, while the *Eucalyptus* plantations (FEP and EG) showed the minimum enrichment. The highest enrichment was DON (10.3 times) for ONE; and TDP (10.7 times) for ND catchment. In general, the differences in enrichment are attributed to high LAI (Leaf Area Index) values in both native forests: the old-growth evergreen forests are multi-stratified and have an understory of high diversity, and particularly in the mixeddeciduous forest the presence of a thick layer of bamboo (*Chusquea quila*), which covered the soil. Our results differing from forested sites in North America and Europe which indicates that the canopies are generally acting as sinks for inorganic-N [33]. Also [40] have reported that NO3 – -N concentrations decreased in stemflow and throughfall relative to precipitation in old-growth forest in North America. However, in a data compilation from 126 European sites with high deposition climate in Scandinavia, Netherlands and Germany, [41] reported that inputs are enhanced by up to 3-5 times in throughfall through addition of dry deposition. On the other hand, our results show that the highest canopy enrichment was DON (dissolved organic nitrogen) especially in both native evergreen and deciduous forests. Also, DON was the most important nutrient fluxes in the native forested catchments, according to the literature [6] that reported that the dominant form of N leaching is dissolved organic nitrogen (DON) in unpolluted forests of southern Chile.

> Annual retention of TDN in native deciduous and evergreen forests was 0.90 and 0.58, and TDP retention was 0.96 and 0.70, respectively. While the exotic *Eucalyptus* plantation there was a net release or loss of 4.79 and 1.44 for TDN and TDP, respectively. Studies in watersheds in the United States [34, 42] reported that thin or porous soils and high infiltration rates have less capacity to retain N. However, in our study, catchments with high infiltration rates, such as evergreen and deciduous forests showed greater N retention than soils with very low infil‐ tration rates, such as *Eucalyptus globulus* plantation. Our results suggests that in native forests, rainfall water was infiltrating and percolating (subsurface flow) exporting less N in contrast to *Eucalyptus* plantation in which as soil has less porosity and infiltration rates due to land use

history. The *Eucalyptus* plantation catchment was cleared (35 years ago) with fire to open areas for grazing animals, and in some areas, for the extraction of wood, and recently (9 years ago) the grassland was replaced by exotic trees (*Eucalyptus globulus*).

Nutrients (TDN and TDP) shows the same behavior in both catchments, their concentration tends to increase as catchment discharge increases. DIN however, showed a different behavior for dry and wet season events. In the native old growth evergreen forest (ONE), DIN lower its concentrations as discharge increased, however in *E. nitens* covered catchment (FEP) increased its concentration. The latter is mostly due to the dilution or the increase of NO3 —N in stream discharge. However, during wet season both catchments showed the same DIN exportation behavior, though FEP had twice as much DIN when compared to ONE.

We are aware that modelling help to unravel and understanding hydrological processes and therefore nutrient exportation occurring within soil catchments. However there are many things to take in to account for, like biota (trees and microorganisms). However, discharge appeared to be a good predictor for TDN and TDP, for both events shown here. This was only seen in FEP, and not in ONE. DIN on the other hand showed poor model fitting. This means that there is still one or several unknowns on the control of DIN exportation during events.

The studies of events provide us with a much detailed perspective of what's happening within the catchment as an ecosystem, either pristine or heavily intervened. The reality is that ecosystems are going to keep "developing", each time with more and more relation to rural and city population. These pristine environments are in great danger and have to be protected from the inhabitants and other anthropic pressures, mostly cattle and land cover change to agricultural lands and exotic species.

Pristine study sites are recognized by being scarce and require a lot of efforts (monetary, time and struggle). In Chile, we have the luxury to have such areas near by some cities, nevertheless it will require more effort to keep it as pristine as possible. The prize for keeping this areas are many, from biodiversity hotspots to be able to unravel some of the black boxes that still exists regarding nutrient exportation and what are the effects of land cover change.

We would like also to address that soil use/cover change history, also plays an important role in N and P retention. Therefore before planting or doing forestry and agricultural activities, soil should be treated in order to enhance nutrient and water retention capabilities.

## **Acknowledgements**

This research was supported by the Fondecyt Project 1120188 (Fondo Nacional de Ciencias). We would like to thank the different owners of the research sites, Mr. Armin Alba, CEFOR (Universidad Austral de Chile), Forestal ANCHILE and Llancahue community for providing the facilities and for collaborating in the monitoring and field work.

## **Author details**

history. The *Eucalyptus* plantation catchment was cleared (35 years ago) with fire to open areas for grazing animals, and in some areas, for the extraction of wood, and recently (9 years ago)

Nutrients (TDN and TDP) shows the same behavior in both catchments, their concentration tends to increase as catchment discharge increases. DIN however, showed a different behavior for dry and wet season events. In the native old growth evergreen forest (ONE), DIN lower its concentrations as discharge increased, however in *E. nitens* covered catchment (FEP) increased

discharge. However, during wet season both catchments showed the same DIN exportation

We are aware that modelling help to unravel and understanding hydrological processes and therefore nutrient exportation occurring within soil catchments. However there are many things to take in to account for, like biota (trees and microorganisms). However, discharge appeared to be a good predictor for TDN and TDP, for both events shown here. This was only seen in FEP, and not in ONE. DIN on the other hand showed poor model fitting. This means that there is still one or several unknowns on the control of DIN exportation during events.

The studies of events provide us with a much detailed perspective of what's happening within the catchment as an ecosystem, either pristine or heavily intervened. The reality is that ecosystems are going to keep "developing", each time with more and more relation to rural and city population. These pristine environments are in great danger and have to be protected from the inhabitants and other anthropic pressures, mostly cattle and land cover change to

Pristine study sites are recognized by being scarce and require a lot of efforts (monetary, time and struggle). In Chile, we have the luxury to have such areas near by some cities, nevertheless it will require more effort to keep it as pristine as possible. The prize for keeping this areas are many, from biodiversity hotspots to be able to unravel some of the black boxes that still exists

We would like also to address that soil use/cover change history, also plays an important role in N and P retention. Therefore before planting or doing forestry and agricultural activities,

This research was supported by the Fondecyt Project 1120188 (Fondo Nacional de Ciencias). We would like to thank the different owners of the research sites, Mr. Armin Alba, CEFOR (Universidad Austral de Chile), Forestal ANCHILE and Llancahue community for providing

regarding nutrient exportation and what are the effects of land cover change.

the facilities and for collaborating in the monitoring and field work.

soil should be treated in order to enhance nutrient and water retention capabilities.

—N in stream

its concentration. The latter is mostly due to the dilution or the increase of NO3

behavior, though FEP had twice as much DIN when compared to ONE.

agricultural lands and exotic species.

**Acknowledgements**

the grassland was replaced by exotic trees (*Eucalyptus globulus*).

348 14 Biodiversity in Ecosystems - Linking Structure and Function

Carlos E Oyarzún1\* and Pedro Hervé-Fernandez2

\*Address all correspondence to: coyarzun@uach.cl

1 Instituto de Ciencias Ambientales y Evolutivas, Facultad de Ciencias, Universidad Austral de Chile, Valdivia, Chile

2 Laboratory of Hydrology and Water Management, Faculty Bioscience Engineering, Uni‐ versity of Ghent, Belgium

## **References**


[20] Lara A, Reyes R, Urrutia R. Bosques Nativos. In: Instituto de Asuntos Públicos, Uni‐ versidad de Chile (eds.), Informe País: Estado del Medio Ambiente en Chile, Santia‐ go, Chile; 2006. p107–139.

[9] Oyarzun C, Godoy R, Sepulveda A. Water and nutrient fluxes in a cool temperate rainforest at the Cordillera de la Costa in southern Chile. Hydrological Processes

[10] Weathers K, Lovett G, Likens G, Caraco N. Cloudwater input of nitrogen to forest ecosystems in southern Chile: forms, fluxes, and sources. Ecosystems 2000; 3:

[11] Weathers K. The importance of cloud and fog in the maintenance of ecosystems.

[12] Oyarzún C, Godoy R, Aracena A, Rutherford P, Deschrijver A. Effects of land use conversion from native forest to exotic plantations on nitrogen and phosphorus re‐ tention in catchments of southern Chile. Water, Air and Soil Pollution 2007; 179:

[13] Armesto JJ, Smith-Ramírez C, Carmona M, Celis-Diez JL, Díaz I, Gaxiola A, Gutiér‐ rez AG, Nuñez-Avila M, Pérez C, Rozzi R. Old-growth temperate rain forests of South America: conservation, plant-animal interactions, and baseline biogeochemical processes. In: Wirth Ch, Gleixner G & Heiman M (eds). Old-growth Forests. Func‐ tion, Fate and Value. Ecological Studies 207, Springer Berlin, Germany; 2009.

[14] Olson D, Dinerstein E. The global 200: a representation approach to conserving the earth´s most biologically valuable ecoregions. Conservation Biology 1998; 12:

[15] Armesto J, Aravena JC, Villagrán C, Pérez C, Parker GG. Bosques templados de la cordillera de la costa. In: Armesto, J.J., Villagrán, C., Arroyo, M.K. (eds.), Ecología de los Bosques Nativos de Chile. Editorial Universitaria, Santiago, Chile; 1996. p199–

[16] Lara A, Villalba R. A 3,620-year temperature record from Fitzroya cuppressoides tree

[17] Nahuelhual L, Donoso P, Lara A, Nuñez D, Oyarzún C, Neira E. Valuing ecosystem services of Chilean temperate rainforest. Environment, Development and Sustaina‐

[18] Nuñez D, Nahuelhual L, Oyarzun C. Forests and water: the value of native temper‐ ate forests in supplying water for human consumption. Ecological Economics 2006;

[19] Echeverría C, Coomes D, Salas J, Rey Benayas JM, Lara A, Newton A. Rapid defores‐ tation and fragmentation of Chilean temperate forests. Biology Conservation 2006;

rings in southern South America. Science 1993; 260: 1104-1106.

Trends in Ecology and Evolution 1999; 14: 214-215.

1998; 12: 1067-1077.

350 16 Biodiversity in Ecosystems - Linking Structure and Function

590-595.

341-350

p367-390.

502-515.

213.

bility 2007; 9: 481-499.

58: 606-616.

130: 481-494.


**Provisional chapter Chapter 14**
