Riparian evapotranspiration modelling: model description and implementation for predicting vegetation spatial distribution in semi-arid environments

• Published in: Ecohydrology Vol. 7; no. 2; pp. 659 - 677
• Main Authors: García-Arias, A., Francés, F., Morales-de la Cruz, M., Real, J., Vallés-Morán, F., Garófano-Gómez, V., Martínez-Capel, F.
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01.04.2014; Wiley Subscription Services, Inc
• Subjects: evapotranspiration modelling; Freshwater; functional types; riparian vegetation; soil moisture; spatial distribution; Spain
• ISSN: 1936-0584; 1936-0592
• DOI: 10.1002/eco.1387

ABSTRACT Biotic and abiotic interactions between the riparian zone and the river determine relevant hydrological processes and exert control over riparian and bordering upland vegetation types. Vegetation growth and development are mainly controlled by water availability on semi‐arid regions, where a moisture gradient determines the transition between the densely vegetated riparian zone and the semi‐arid upland. To reproduce this spatial distribution, a mathematical model named RibAV is presented. Its conceptualization is based on the main ecohydrological modelling approaches and field expertise. The implementation of RibAV that is proposed in this paper allows the simulation of the distribution of three plant functional types [herbaceous riparian vegetation (HRV), woody riparian vegetation (WRV) and terrestrial vegetation (TV)] within the riparian zone. An evapotranspiration index (Eidx) obtained through RibAV is used as a criterion for plant absence/presence prediction. Two permanent river reaches of semi‐arid Mediterranean basins, the Terde reach (Mijares River, Spain) and the Lorcha reach (Serpis River, Spain), have been selected as case studies for the calibration and validation of the model, respectively. Several criteria based on the confusion matrix were used to analyse the efficiency of RibAV on the prediction of plant distribution. The satisfactory performance of the model establishing the distribution of the riparian vegetation types and the limit between this zone and the bordering upland are demonstrated in this paper; the strength of the Eidx to classify plant functional types in riparian semi‐arid environments is additionally proven. Copyright © 2013 John Wiley & Sons, Ltd.