**4. Flood delimitation mapping of the Tavira urban area**

#### **4.1. Introduction and objective**

This case study illustrates the use of GIS as a tool to establish hydrologic regional parameters for urban flood mapping purposes. Cartographic elements, hydrological and hydraulic models, and boundary conditions used to establish the maximum flood levels of a 10- and a 100-years return period flood and a 100-years climate change scenario are described. Cartographic information was completed by *in-situ* measurements and observations.

Hydrologic regional parameters are extensively used in flood simulation, since drainage basins are characterized by natural variability in land-surface features (*e.g.*, Wooldridge and Kalma, 2001). Prasad (1997) refers that the improved accuracy of GIS-based hydrologic simulation comes from the capability that these models have to integrate hydrologic regional parameters; updating or modifying GIS data to study the impact of changes in a drainage basin (*e.g.*, land use) becomes a relatively easy task.

This application is focused on the simulation of fluvial-originated urban flooding. The area selected for this study is the town of Tavira. This town is situated in the southernmost region of Portugal – Algarve. The Séqua/Gilão River crosses throughout the Tavira urban area until it flows into the Ria Formosa coastal lagoon. As the Séqua/Gilão River is intrinsically connected with the urban fabric, an overtopping of the margins always has negative consequences to people and assets. An example of a severe flood event was the 3rd December 1989 flood which caused extensive damage in the city.

#### **4.2. Study area**

Tavira (12,000 inhabitants) is one of the southernmost towns of mainland Portugal. The town origins come back since about 2,000 BC and from these years until the half of the last century Tavira has had in agriculture and fishing its major economical activities. Like many areas in the Algarve region, in these last decades, tourism has gained a sound importance in the local economy and lifestyle. The Séqua/Gilão River, which crosses the entire town, is crossed by 6 bridges.

GIS-Based Models as Tools for Environmental Issues: Applications in the South of Portugal 267

**Sub-basins Alportel Asseca Fornalha Séqua** 

The Séqua/Gilão drainage basin has an area of 227 km2 immediately upstream of Tavira. This drainage basin can be divided into 4 sub-basins, namely: Alportel, Asseca, Fornalha

Sub-basin area (km2) 93.12 61.35 40.06 32.31 Main river length (km) 49.07 19.77 15.06 14.25 Equivalent slope (m/km) 6.5 9.4 17.6 5.9 Time of Concentration (h) 9.2 4.0 2.5 3.7 Lag-time (h) 5.5 2.4 1.5 2.2

**Table 9.** Major hydrological and physical characteristics of the drainage sub-basins and rivers of the

**Figure 6.** Drainage basin of the Séqua/Gilão River immediately upstream of Tavira. The 4 sub-basins

floods. Adjusted values (CN III) used in this study are shown in Figure 7.

According to the Portuguese Soils Map (SROA, 1970) most of the soils of the Séqua/Gilão drainage basin may be classified as "Ex - Lithosols of xeric regime climate, schists or greywackes" (Cardoso, 1965). Soils properties were defined using field tests data (Koop *et al.*, 1989). The soils hydrologic groups were defined (*e.g.*, Lencastre and Franco, 1992) as B, C and D (US Soil Conservation Service Curve Number method). Group D was by far the most common in the drainage basin. Soil use was acquired via the Corine Land Cover chart and largely consists of forests and farmlands. According to the Curve Number method, information on the soil types and uses were combined to obtain the CN II values, to which soil types and uses determine the relationship between rainfall and effective rainfall for soil moisture conditions between the wilting point and field capacity. However, in flood-related studies, the soil moisture conditions should correspond to soils with high water content which are able to reach or surpass field capacity, a situation that leads to the origin of larger

The design hyetographs were obtained by using Intensity-Duration-Frequency curves referred in the Portuguese Law (DR 23/95 of 23rd August) and the "alternating block

and Séqua (Figure 6). Characteristics of the sub-basins are shown in Table 9.

*4.3.1. Hydrologic model* 

Séqua/Gilão River.

are identified.

**Characteristics of the drainage sub-basins and rivers** 

This study characterizes the whole river basin of the Rio Séqua/Gilão with length of the river valley of 9.5 km from the outlet. The drainage basin of the Séqua/Gilão River considered for this study is located immediately upstream of the northern limit of Tavira's urban area (top left of Figure 5).

## **4.3. Materials and methods**

This section presents the hydrologic and hydrodynamic models used to determine the peak flows and the delimitation of the flooded areas. Particular attention was given to obtaining the hydrologic parameters related to soil use and type (CN values; Brunner, 2006). The flood delimitation methodology and the verification of the modelled flood – by comparison with observed data – are also presented.

**Figure 5.** Urban area of Tavira and the Séqua Gilão River, flowing from left to right.

#### *4.3.1. Hydrologic model*

266 Cartography – A Tool for Spatial Analysis

entire town, is crossed by 6 bridges.

**4.3. Materials and methods** 

observed data – are also presented.

Tavira (12,000 inhabitants) is one of the southernmost towns of mainland Portugal. The town origins come back since about 2,000 BC and from these years until the half of the last century Tavira has had in agriculture and fishing its major economical activities. Like many areas in the Algarve region, in these last decades, tourism has gained a sound importance in the local economy and lifestyle. The Séqua/Gilão River, which crosses the

This study characterizes the whole river basin of the Rio Séqua/Gilão with length of the river valley of 9.5 km from the outlet. The drainage basin of the Séqua/Gilão River considered for this study is located immediately upstream of the northern limit of Tavira's urban area (top

This section presents the hydrologic and hydrodynamic models used to determine the peak flows and the delimitation of the flooded areas. Particular attention was given to obtaining the hydrologic parameters related to soil use and type (CN values; Brunner, 2006). The flood delimitation methodology and the verification of the modelled flood – by comparison with

**Figure 5.** Urban area of Tavira and the Séqua Gilão River, flowing from left to right.

**4.2. Study area** 

left of Figure 5).

The Séqua/Gilão drainage basin has an area of 227 km2 immediately upstream of Tavira. This drainage basin can be divided into 4 sub-basins, namely: Alportel, Asseca, Fornalha and Séqua (Figure 6). Characteristics of the sub-basins are shown in Table 9.


**Table 9.** Major hydrological and physical characteristics of the drainage sub-basins and rivers of the Séqua/Gilão River.

**Figure 6.** Drainage basin of the Séqua/Gilão River immediately upstream of Tavira. The 4 sub-basins are identified.

According to the Portuguese Soils Map (SROA, 1970) most of the soils of the Séqua/Gilão drainage basin may be classified as "Ex - Lithosols of xeric regime climate, schists or greywackes" (Cardoso, 1965). Soils properties were defined using field tests data (Koop *et al.*, 1989). The soils hydrologic groups were defined (*e.g.*, Lencastre and Franco, 1992) as B, C and D (US Soil Conservation Service Curve Number method). Group D was by far the most common in the drainage basin. Soil use was acquired via the Corine Land Cover chart and largely consists of forests and farmlands. According to the Curve Number method, information on the soil types and uses were combined to obtain the CN II values, to which soil types and uses determine the relationship between rainfall and effective rainfall for soil moisture conditions between the wilting point and field capacity. However, in flood-related studies, the soil moisture conditions should correspond to soils with high water content which are able to reach or surpass field capacity, a situation that leads to the origin of larger floods. Adjusted values (CN III) used in this study are shown in Figure 7.

The design hyetographs were obtained by using Intensity-Duration-Frequency curves referred in the Portuguese Law (DR 23/95 of 23rd August) and the "alternating block

method" (Chow, 1988). Kirpich formula (*e.g.*, Guo, 2006; de Lima, 2010) was used do estimate the Time of Concentration of each of the sub-basins; the Lag-times were approximated as 60% of the latter (Table 9).

GIS-Based Models as Tools for Environmental Issues: Applications in the South of Portugal 269

The HEC–RAS model (Hydraulic Engineering Center – River Analysis System) was used to define, for the full extension of the Séqua/Gilão River, the maximum flood levels to be expected for 10- and 100-years recurrence periods. In the simulation it was assumed (i) a 1D unsteady flow, (ii) inflow in the upstream boundary condition is defined by the previously obtained design hydrographs, (iii) water level at the downstream boundary condition is defined by expected spring tide levels at Faro bar and an additional climate change scenario with a mean sea rise of 0.91 m was also used (mean sea rise value was imposed by national authorities), (iv) flow resistance is approximated by the Manning-Strickler equation, (v) densely urbanized areas are considered as non-effective flow areas, *i.e.*, where water overflows and returns in the same river section and (iv) head loss in hydraulic structures (*e.g.*, bridges) is due both to the head loss during flow along the hydraulic structure and the

The main flow line was discretized by cross sections (STs) of the river, based in topographic and bathymetric surveys. These STs were set where the river geometry showed important changes and near hydraulic structures. Geometrical information of the latter was retrieved from the structures final drawings and *in situ* measurements made for this purpose. Riverbed material was classified by local observations (Chow, 1959). Maximum flood levels in the Tavira urban area were obtained for the worst case scenario, *i.e.*, when the downstream-moving river flood wave overlaps the upstream-moving tide wave. This case scenario takes place when the river peak flow reaches the upstream boundary cross section

1 hour before the tidal high water occurs at the downstream boundary cross section.

define the 10- and 100-years recurrence period flood-affected areas:

Topographical map of Tavira with contour lines every 10 m;

*4.3.4. Comparison of the model results with observed flood levels* 

Portuguese Army topographical maps 1:25,000 and ortophotomaps;

The following cartographical data and the hydrodynamic model results were combined to

Bathymetric survey of the Séqua/Gilão riverbed and the Tavira bar, Quatro Águas bar

The flood-affected areas were delimited for the 10- and 100-years return period and the 100 years climate change scenario by retrieving, for each of the cross sections, the maximum flood levels from the hydrodynamic model. After completion of this process, the areas above the maximum flood which were within the flood delimitated area (island areas) were

In the night of the 3rd of December 1989 a severe flood occurred in Tavira with 120 mm of daily rainfall registered during that day at the São Brás de Alportel meteorological station.

*4.3.2. Hydrodynamic model* 

localized head loss.

*4.3.3. Flood delimitation* 

and Gilão River.

trimmed out.

**Figure 7.** Soil Conservation Service CN III values for the drainage basin of the Séqua/Gilão River.

The information above served as an input to the HEC–HMS model (Brunner, 2006). Data for the drainage basin watershed were introduced into the model, differentiating the four sub-basins and their topological connections, the reaches and the final section of the drainage basin (sink). Resulting Séqua/Gilão River basin model is shown schematically (Figure 8).

**Figure 8.** Topological scheme of the Séqua/Gilão River drainage basin.

The resulting design hydrographs produced peak flows for the 10- and 100-years return period, respectively, of 520.6 m3/s and 928.0 m3/s.

## *4.3.2. Hydrodynamic model*

268 Cartography – A Tool for Spatial Analysis

(Figure 8).

approximated as 60% of the latter (Table 9).

method" (Chow, 1988). Kirpich formula (*e.g.*, Guo, 2006; de Lima, 2010) was used do estimate the Time of Concentration of each of the sub-basins; the Lag-times were

**Figure 7.** Soil Conservation Service CN III values for the drainage basin of the Séqua/Gilão River.

**Figure 8.** Topological scheme of the Séqua/Gilão River drainage basin.

period, respectively, of 520.6 m3/s and 928.0 m3/s.

The information above served as an input to the HEC–HMS model (Brunner, 2006). Data for the drainage basin watershed were introduced into the model, differentiating the four sub-basins and their topological connections, the reaches and the final section of the drainage basin (sink). Resulting Séqua/Gilão River basin model is shown schematically

The resulting design hydrographs produced peak flows for the 10- and 100-years return

The HEC–RAS model (Hydraulic Engineering Center – River Analysis System) was used to define, for the full extension of the Séqua/Gilão River, the maximum flood levels to be expected for 10- and 100-years recurrence periods. In the simulation it was assumed (i) a 1D unsteady flow, (ii) inflow in the upstream boundary condition is defined by the previously obtained design hydrographs, (iii) water level at the downstream boundary condition is defined by expected spring tide levels at Faro bar and an additional climate change scenario with a mean sea rise of 0.91 m was also used (mean sea rise value was imposed by national authorities), (iv) flow resistance is approximated by the Manning-Strickler equation, (v) densely urbanized areas are considered as non-effective flow areas, *i.e.*, where water overflows and returns in the same river section and (iv) head loss in hydraulic structures (*e.g.*, bridges) is due both to the head loss during flow along the hydraulic structure and the localized head loss.

The main flow line was discretized by cross sections (STs) of the river, based in topographic and bathymetric surveys. These STs were set where the river geometry showed important changes and near hydraulic structures. Geometrical information of the latter was retrieved from the structures final drawings and *in situ* measurements made for this purpose. Riverbed material was classified by local observations (Chow, 1959). Maximum flood levels in the Tavira urban area were obtained for the worst case scenario, *i.e.*, when the downstream-moving river flood wave overlaps the upstream-moving tide wave. This case scenario takes place when the river peak flow reaches the upstream boundary cross section 1 hour before the tidal high water occurs at the downstream boundary cross section.

#### *4.3.3. Flood delimitation*

The following cartographical data and the hydrodynamic model results were combined to define the 10- and 100-years recurrence period flood-affected areas:


The flood-affected areas were delimited for the 10- and 100-years return period and the 100 years climate change scenario by retrieving, for each of the cross sections, the maximum flood levels from the hydrodynamic model. After completion of this process, the areas above the maximum flood which were within the flood delimitated area (island areas) were trimmed out.

#### *4.3.4. Comparison of the model results with observed flood levels*

In the night of the 3rd of December 1989 a severe flood occurred in Tavira with 120 mm of daily rainfall registered during that day at the São Brás de Alportel meteorological station.

Making use of an amateur video from that night and from flood level inscriptions still visible on some walls, some flood-affected locations were identified (Figure 9). Since it was possible to recognize from the video the maximum flood levels, by means of a set of *in-situ* measurements the depth of water in those locations was estimated (Table 10).

GIS-Based Models as Tools for Environmental Issues: Applications in the South of Portugal 271

**Figure 10.** Flood delimitation for the Tavira urban area. Blue and magenta areas represent, respectively,

This application shows an example of how GIS-based soil data may be used as an input for flood-modelling purposes. In this example HEC-HMS (hydrological model) and HEC-RAS (hydraulic model) were used to obtain maximum flood levels for 10- and 100-years return

This work aimed to illustrate how GIS-based models can be used as tools for environmental studies through three case studies in the south of mainland Portugal. The first dealt with the problem of geo-form classification in the Ria Formosa estuary. The second focused on using DEM/DTM-based climate models to obtain and analyze isohyetal maps and to identify rainfall distribution influence on water erosion at the Serra de Grândola. In the third application, GIS-based models were used to determine hydrological regional parameters for urban flood mapping purposes (this last application focused in the Séqua/Gilão River and the city of Tavira). The applications allowed demonstrating the versatility and usefulness of

GIS-based models when used to solve environmental issues.

*University of Algarve, Campus da Penha, Faro, Portugal* 

*IMAR – Marine and Environmental Research Centre, Department of Civil Engineering,* 

the 10- and 100-years return period.

**5. General conclusions** 

**4.5. Conclusions** 

**Author details** 

Jorge M. G. P. Isidoro

periods.


**Table 10.** Comparison between the simulation results and the observed flood depths attained on the 3rd of December 1989 flood event.

**Figure 9.** Location of places used to compare the simulation results and the observed flood depths attained on the 3rd of December 1989 flood event.

#### **4.4. Results and discussion**

The 10- and 100- year flood area maps of the Tavira urban area show that a significant part of the urban area which is adjacent to the Séqua/Gilão River is within the flood-affected perimeter (Figure 10). The City's centre is severely affected both by the simultaneous occurrence of high spring tide and the 10- or the 100-year rainfall events.

It is clearly visible that after the ST11 section the flood area expands widely. This is because ST11 represents a heritage bridge which causes significant obstruction to the river channel. A hydraulic gradient is formed by this bridge thus allowing water to overflow the river channel.

GIS-Based Models as Tools for Environmental Issues: Applications in the South of Portugal 271

**Figure 10.** Flood delimitation for the Tavira urban area. Blue and magenta areas represent, respectively, the 10- and 100-years return period.

#### **4.5. Conclusions**

270 Cartography – A Tool for Spatial Analysis

of December 1989 flood event.

attained on the 3rd of December 1989 flood event.

**4.4. Results and discussion** 

channel.

Making use of an amateur video from that night and from flood level inscriptions still visible on some walls, some flood-affected locations were identified (Figure 9). Since it was possible to recognize from the video the maximum flood levels, by means of a set of *in-situ*

**Location (point) Observed flood depth (m) Simulated flood depth (m) A** 2.5 2.9 **C** 4.4 4.3 **D** 4.4 4.3 **H** 2.7 3.1 **K** 4.1 4.1 **M** 4.4 4.3 **Table 10.** Comparison between the simulation results and the observed flood depths attained on the 3rd

**Figure 9.** Location of places used to compare the simulation results and the observed flood depths

The 10- and 100- year flood area maps of the Tavira urban area show that a significant part of the urban area which is adjacent to the Séqua/Gilão River is within the flood-affected perimeter (Figure 10). The City's centre is severely affected both by the simultaneous

It is clearly visible that after the ST11 section the flood area expands widely. This is because ST11 represents a heritage bridge which causes significant obstruction to the river channel. A hydraulic gradient is formed by this bridge thus allowing water to overflow the river

occurrence of high spring tide and the 10- or the 100-year rainfall events.

measurements the depth of water in those locations was estimated (Table 10).

This application shows an example of how GIS-based soil data may be used as an input for flood-modelling purposes. In this example HEC-HMS (hydrological model) and HEC-RAS (hydraulic model) were used to obtain maximum flood levels for 10- and 100-years return periods.
