Assessment of brown trout habitat suitability in the Jucar River Basin (SPAIN): Comparison of data-driven approaches with fuzzy-logic models and univariate suitability curves

Abstract

The implementation of the Water Framework Directive implies the determination of an environmental flow (E-flow) in each running water body. In Spain, many of the minimum flow assessments were determined with the physical habitat simulation system based on univariate habitat suitability curves. Multivariate habitat suitability models, widely applied in habitat assessment, are potentially more accurate than univariate suitability models. This article analyses the microhabitat selection by medium-sized (10–20 cm) brown trout (Salmo trutta fario) in three streams of the Jucar River Basin District (eastern Iberian Peninsula). The data were collected with an equal effort sampling approach. Univariate habitat suitability curves were built with a data-driven process for depth, mean velocity and substrate classes; three types of data-driven fuzzy models were generated with the FISH software: two models of presence–absence and a model of abundance. FISH applies a hill-climbing algorithm to optimize the fuzzy rules. A hydraulic model was calibrated with the tool River-2D in a segment of the Cabriel River (Jucar River Basin). The fuzzy-logic models and three methods to produce a suitability index from the three univariate curves were applied to evaluate the river habitat in the tool CASiMiR©. The comparison of results was based on the spatial arrangement of habitat suitability and the curves of weighted usable area versus discharge. The differences were relevant in different aspects, e.g. in the estimated minimum environmental flow according to the Spanish legal norm for hydrological planning. This work demonstrates the impact of the model's selection on the habitat suitability modelling and the assessment of environmental flows, based on an objective data-driven procedure; the conclusions are important for the water management in the Jucar River Basin and other river systems in Europe, where the environmental flows are a keystone for the achievement of the goals established in the European Water Framework Directive.

Introduction

In the Iberian Peninsula the prediction of the future average rainfall indicates a decreasing trend (Rodrigo and Trigo, 2007), which will cause a decrease of the ecosystem services (Schröter et al., 2005). For instance the changes in the flow regime will produce a decrease in the recruitment and survival of brown trout populations (Jonsson and Jonsson, 2009). Hence, new conflicts will appear between human needs and the conservation of aquatic species. In Spain, as it was established in the norm for hydrological planning (Instrucción de Planificación Hidrológica, 2008), the environmental flows (hereafter E-flows) were estimated with the application of the physical habitat simulation, a technique originally developed in the frame of the Instream Flow Incremental Methodology (IFIM; Bovee, 1982), involving hydraulic simulation and habitat evaluation with habitat suitability models, frequently in the form of univariate habitat suitability curves (HSCs). These curves relate the simulated hydraulic variables and other habitat variables (commonly substrate and cover) with the index of habitat suitability (from 0, unacceptable for the aquatic species, to 1, optimal); the indices for these variables are combined to produce the index called Weighted Usable Area (WUA) (Bovee and Cochnauer, 1977), which weights habitat suitability and area in a river reach, for a given flow rate. The relationship WUA versus discharge and the habitat time series analysis are fundamental information to support decisions on water management projects in the methodological frame of the IFIM.

According to the Spanish IPH the minimum flow should be selected within the range of 50–80% of the maximum WUA. This legal requirement remarks the importance of a reliable knowledge on fish habitat suitability and its variation with discharge. The enhancement of this knowledge is fundamental to improve habitat predictions under unsurveyed flows (Drew et al., 2011). Several authors have suggested that considering each hydraulic variable independently in a suitability curve may be questionable (Lambert and Hanson, 1989); to deal with this limitation, the multivariate approach has increased in popularity among researchers (De Pauw et al., 2006).

In the context of the Mediterranean rivers, the studies on fish habitat suitability models are scarce; this may be due to the high percentage of endemisms and their limited distribution area (Ferreira et al., 2007). In the Iberian Peninsula, most of the studies have produced univariate models, in PhD thesis (Martínez Capel, 2000, Martínez-Capel and García de Jalón, 1999) or in grey literature as technical reports (e.g. Martínez Capel et al., 2011). Only a few multivariate models are available in this context; for Iberian fish species there are two examples at the mesohabitat scale, the generalized additive models (GAMs) for the endangered Jucar nase, Parachondrostoma arrigonis (Costa et al., 2011) and a fuzzy-logic model for brown trout (Mouton et al., 2011). At the microhabitat scale, there are Resource Selection Functions (RSFs) (Ayllón et al., 2009) and an expert-knowledge fuzzy model (Magdaleno Mas and Martínez Romero, 2005), both for brown trout. Regarding the latter study, some authors questioned the reliability and robustness of habitat assessment based on expert-knowledge habitat models because some inconsistencies among experts were found appraising the suitability of given conditions (Acreman and Dunbar, 2004). Data-driven fuzzy models have been successful in complementing expert knowledge-based models, improving their reliability. In general, fuzzy-logic models take into account the inherent uncertainty of ecological variables and enable expressing non-linear relations between ecological variables in a transparent manner; they have been acknowledged as a suitable approach for species distribution modelling and habitat suitability assessments (Mouton et al., 2011, Mouton et al., 2008a, Mouton et al., 2008b).

However, there are no experimental studies comparing E-flow assessments based on HSCs and based on data-driven fuzzy models. This paper aims to fill this knowledge gap by the comparison of both techniques in an Iberian river.

Three types of data-driven fuzzy models were generated: Presence–Absence (P/A), Overpredictive Presence–Absence (P/A-O) and Abundance (Ab) models, with minimum intervention of the expert knowledge for an objective comparison. Presence–Absence models predict the degree of fulfilment of the habitat conditions where the fish were observed. These models determine the degree of habitat suitability for the target species but they do not provide information in terms of the potential number of fish under a given habitat configuration. The Abundance models have similar output but provide information in terms of the number of individuals potentially present under given conditions. The outputs of all these models range between zero and one; this fact allows the modeller to compare the results, but the meanings of the outputs are different. Therefore, for P/A models, one means presence and zero means absence. In Ab models the output is the standardized fish abundance, thus one is the maximum number of fish observed and zero means that no fish were observed. The difference between P/A models and P/A-O models lies in the frequency of presence, being more frequent in the overpredictive models.

The fuzzy models were used, together with HSCs, to assess the habitat in a nearly pristine Iberian river for a wide range of flows. The habitat assessments derived from the HSCs and from the fuzzy models were compared spatially by calculating their percentage of agreement. In addition, the WUA-flow curves were generated and compared. Finally, the legal implication of selecting any of the generated models was discussed.