Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Pays Bas → 2020 → Remotely sensed and hydrograph-derived early warning signals of desertification in degrading Spanish catchments

Utrecht University (2020)

Remotely sensed and hydrograph-derived early warning signals of desertification in degrading Spanish catchments

Veldhuijzen, P.

Titre : Remotely sensed and hydrograph-derived early warning signals of desertification in degrading Spanish catchments

Auteur : Veldhuijzen, P.

Université de soutenance : Utrecht University

Grade : Master of Science (MS) 2020

Résumé _ Past research suggested the presence of early warning signals (EWS) based on higher-order statistics of system state variables, and that these can contribute to the prediction of critical transitions of ecosystems. Modelling studies have shown that the spatial and temporal variability of biomass are promising EWSs for desertifying semi-arid systems, but only limited research has been conducted to real-world systems. The strong theoretical connections between vegetation, land surface temperature and river discharge suggest that EWSs may be present for each of these system state variables (SSVs). In this study, we assessed the relations between spatial and temporal EWSs and the SSVs by analyzing multidecadal timeseries of remotely sensed data (NDVI, representing vegetation and land surface temperature) and of river discharge data for several (potentially) degrading catchments in Spain. We show that the trends over time of (catchment-averaged) SSVs and EWSs are often related to each other. Although some combinations of SSVs and EWSs show a different relation to each other than expected based on previous research, there are often dominant relations in a majority of the studied catchments. Particularly spatial variability-based EWSs tend to often show a clear relation (either similar or opposite) to SSVs. Temporal variability-based EWSs show less often a dominant trend over time, except for EWSs derived from river discharge. We show that temporal and spatial EWSs of the same SSV do not always show a clear similarity, implying that the previously suggested similar behavior of spatial and temporal EWSs can often not be observed practice. Particularly the biomass and LST and their derived EWSs often show a covariation, which suggests that EWSs may indeed be found in various system components ; river discharge and its variability are likely strongly influenced by precipitation changes as well. Furthermore, we found differences between different types of variability measures used for EWSs (namely coefficient of variation and the standard deviation) ; practical application of EWSs requires research to which measure is most suitable. Although the observed relations between SSVs and EWSs are sometimes else than previously expected, the presence of these relations suggests that SSVs and EWSs can contain additional information with respect to predicting critical transitions towards desertified systems

Présentation

Page publiée le 17 avril 2021