**6. The functional role of reef fish**

A common feature for both collection periods is the low δ15N values of macroalgae. During winter-2006, with the exception of Chankanaab reef at PNAC (δ<sup>15</sup>N of 5.0 ± 2.2 ‰), the δ15N of macroalgae is less than 3 ‰ and is markedly smaller than the isotopic value of nitrate (δ15NO3) (Figure8a). As in the winter of 2006, the average δ15N of macroalgae during the summer of 2007

It is noteworthy to highlight the results obtained from Chankanaab reef (at PNAC) which shows the greatest contrast in the average δ15N of macroalgae between sampling seasons: while the average δ15N of macroalgae was the most positive of all sampling stations during the winter of 2006 (Figure 8a), during the summer of 2007 Chankanaab reef presented the most negative δ15N values (-0.34 ± 1.43 ‰) of all the studied sites (Figure 8b). One factor that may explain this discrepancy is the difference in species composition between sampling sites. While in the winter of 2006 the δ15N was measured in *Dyctiota* spp. and *Penicilluspyri‐ formis*, during summer 2007 the species analyzed were *Lobophoravariegata*, *Dictyosphaera cav‐ ernous,Anadyomenestellata*and *Ulotrix* spp. The results for species indicates that the δ15N of *Dyctiota* spp. and *P. pyriformis* were characterized by positive values in both sampling peri‐ ods, while the species *L. variegata* always showed low δ15N values, including *A. stellata* whose measured δ15N was the most negative (-1.67 and -1.42 ‰ for Chankanaab and Dalila,

There are several alternatives to explain the low δ<sup>15</sup>N values of macroalgae in this re‐ gion. N2 fixation in coral reefs is regarded as a major component of the nitrogen cycle that provides new nitrogen for these ecosystems. It has been estimated that fixed nitro‐ gen can supply from one quarter to one half of the nitrogen requirements for the pri‐ mary producers in these oligotrophic environments [30]. It has been shown in coral reef areas where nitrogen fixation is predominant that δ15N value of macroalgae is close to 0 ‰, or even negative [31,32]. In agreement, the range of δ<sup>15</sup>N values measured in the dif‐ ferent species of macroalgae (<0 to <2.5 ‰) in our study, indicates that N2 fixation may be playing an important role in fulfilling the macrophytes' nitrogen demand. This con‐ clusion is also supported by [25] who found that the δ<sup>15</sup>N values of 1.9 ‰ in the sea‐ grass *Thalassia testudinum,* in Puerto Morelos lagoon, were the result of nitrogen fixation. Alternatively, another possibility is the low concentration of dissolved inorganic nitrogen (DIN) in the studied reefs. Different studies have shown a positive relationship between the δ15NO3 *versus* nitrate concentration [NO3], both in temperate and tropical areas. It was found in a Massachusetts estuary that the δ15NO3 values approach 0 ‰ when the nitrate concentra‐ tion is reduced to levels <1 uM [33]. Similarly, other results obtained by [34]from the Mexi‐ can Caribbean reported that δ15NO3 decreases linearly with the concentration of nitrate in the water column.Thus, consideringthe low DIN concentrations that characterizeour study area it would be expected that the δ<sup>15</sup>N-DIN available for macroalgae should be character‐

Lastly, changes in the proportion of the different nitrogen species available for photosyn‐ thesis could explain the lower values in δ15NO3. In our study, ammonia accounts for 60 to 76% of the DIN levels. This implies that ammonium, but not nitrate, could be the main source of nitrogen to seaweeds in the region. Hence, the δ15N of macroalgae largely

varied from ~ 0.5 to 2.5 ‰, being always below the average δ15NO3 (Figure 8b).

respectively) in both sampling periods.

42 Environmental Change and Sustainability

ized by lower values.

Fishes are particularly recognized for their role as the main drivers of energy flow in coral reefs and can be separated into two major functional groups: the grazers or herbivores that regulate the abundance and community structure of algae; and corallivorous that selectively feed on coral tissue. Its importance, in addition to controlling the population of primary pro‐ ducers is that their feeding activity promotes biodiversity, since consumption of algae and coral leaves available space for colonization of new individuals or species [35]. Another im‐ portant group are the predators as they play an important role in the ecosystem because they occupy the highest level in food webs, and from this position they regulate the organ‐ isms that are in the lower trophic levels. Additionally, they connect the dynamics of other communities and ecosystems that are apparently distinct as they often travel long distances. It has been suggested that the predator's ability to travel great distances in response to changes in prey abundance is important for maintaining stability of food webs [36]. Conse‐ quently, the loss of top predators could destabilize ecosystems through a chain reaction that eventually propagate down through the food web. With this in mind, the biomass of herbiv‐ orous and predatory fish have been calculated in order to determine whether there is spatial variability in these functional groups.

There is a close link between the fish community and benthic components, as any change in the structure of one of these communities has an effect on the structure of the other. [37] pointed out that in quantitative studies during the 80's reported high densities of predators, like sharks and groupers, associated with areas of high coral cover. Currently, the presence of large top predators is rare and is considered that the decrease in the abundance of this group has strongly affected the trophic flow patterns in reef communities [38]. In addition, the decrease in coral cover in coral reefs has been related to the decrease in the abundance and diversity of reef fishes [39].

In order to gain a better understanding of the role of fishes in our study area, we analyzed the community structure of reef fish sampled at PNAPM and PNAC; for PNIMCN we only had data for Manchones reef. By applying a similarity analysis (ANOSIM) to the abundance data, significant differences in the structure of the fish community were detected in these lo‐ cations (R = 0.442, *p* < 0.0001). Paired tests showed significant differences between all loca‐ tions, except Yucab and Paraiso Reefs (*p* = 0.09), and Yucab and Paso Cedral reefs (*p* = 0.23); also Dalila and Colombia do not show significant differences (*p* = 0.20). Using cluster analy‐ sis, based on a similarity matrix generated using the Bray-Curtis index, can be seen again the grouping of localities from each national park, leaving the Manchones reef alone as an isolated entity (Figure 9).

**Figure 9.** Cluster analysis of the reef fish community in northern Quintana Roo (2007 data).

The differences in the structure of the fish community, can be as well associated with struc‐ tural differences in the benthic composition of the reefs, as the decline in coral cover reduces the structural complexity of the reef and therefore the space available for shelter and feeding [39] (Figure 10).

**Figure 10.** Trends in fish species richness and coral cover in the three national coral reef parks of Northern Quintana Roo, Mexico.

As to the relative abundance of trophic guilds, we can observe that higher predators gener‐ ally have little relative abundance or simply not recorded during surveys, being Radio Pi‐ rate reef in Puerto Morelos, as well as Dalila and Colombia reefs in Cozumel were the only coral reefs where this trophic guild was recorded. The relative abundance of carnivores is up 50% Chankana'ab and Paso del Cedral reefs, however this trophic group also character‐ ized by small fish whose dietary components include small invertebrates. Another group that has a high relative abundance of 60% and 50%, in Colombia and Dalila reefs respective‐ ly, are the herbivores (Figure 11).

**Figure 9.** Cluster analysis of the reef fish community in northern Quintana Roo (2007 data).

[39] (Figure 10).

44 Environmental Change and Sustainability

Roo, Mexico.

The differences in the structure of the fish community, can be as well associated with struc‐ tural differences in the benthic composition of the reefs, as the decline in coral cover reduces the structural complexity of the reef and therefore the space available for shelter and feeding

**Figure 10.** Trends in fish species richness and coral cover in the three national coral reef parks of Northern Quintana

**Figure 11.** Relative abundance of the trophic guilds of the fish community associated with the coral reefs of northern Quintana Roo (2007 data).

Using the information of the size structure of three reefs (Manchones, Bonanza and Radio Pira‐ ta reefs) we calculated the relative biomass of the different trophic guilds using length – weight parameters for each species available at fish base [40]. This information allows assessing the de‐ gree of disturbance in the communities of each reef studied. In stable conditions or of low-dis‐ turbance, the dominant species of large size and longevity (K-strategists) are dominant in biomass and have low abundance. There are also present in the communities r-strategy spe‐ cies, opportunistic species with a short lifetime that are dominate in abundance but have low biomass contribution. When a community is disturbed, K-strategy species are usually not fa‐ vored and opportunistic species increase in numbers and biomass [19].

For Radio Pirate and Bonanza reefs we can see that the relative biomass of carnivores is high (>50% of the biomass), however, the relative biomass of larger predators is virtually nonexis‐ tent. In the case of Manchones reefs, the herbivores contribute in greater proportion to the biomass (Figure 12). Such inverted biomass pyramids of fish have been reported in coral reefs characterized by low coral cover. To maintain these typesof biomass pyramids, howev‐ er, a high primary production is required [37].

**Figure 12.** Relative biomass of trophic guilds in reef fish communities of northern Quintana Roo coral reefs (2007 data).

Additionally, a comparison analysis chart of abundance/biomass (Abundance / Biomass Comparison - ABC plots) was performed. This method presents a statistical test (W) that represents the abundance over biomass in a range of -1 to +1; if the statistical test generates a +1, biomass dominates over abundance and represents a system with no impact. On the con‐ trary, when the result is -1, abundance dominates over biomass indicating that the system has been highly impacted. Values near zero indicate an intermediate disturbance [19].

In the case of reefs for which we had data on both, abundance and biomass, we obtained W = 0.146, indicating that the fish communities of these reefs are under a scenario of intermedi‐ ate disturbance (Figure 13) that may result either from the degradation of their habitat or the effect of fishing and poaching that take place in these reef locations [41-43]. This analysis al‐ low us to conclude that It is essential to conduct a monitoring program of the fish communi‐ ty structure, and their biomass, to better assess their status and how they contribute to the functioning of these ecosystems.

**Figure 13.** ABC-plot of reef fish for three coral reefs (Manchones, Bonanza and Pirate Radio) of northern Quintana Roo (2007 data).
