Assessing the drivers of biodiversity in Madagascar by quantifying its hydrologic properties at the watershed scale

• Multi-year remote-sensing observations used to quantify hydrologic properties
• Investigate the intra-annual responses of Malagasy basins to precipitation anomaly
• Evaluate the drought-response hypothesis of Wilmé et al. (2006)
• Provide insight into the hydro-geomorphologic drivers of biodiversity patterns

Motivated by a theory developed by Wilmé et al. (2006) according to which, watersheds of Madagascar with headwaters at high altitude respond differently to drought from those with headwaters confined to relatively low elevations, with possibly profound effects on the biodiversity patterns of the island, we analyzed multi-year basin-specific observations of soil moisture and vegetation water content (derived from NRL's WindSat radiometer) and their response to precipitation departures (derived from TRMM 3B42 V7) from its local mean. These datasets were analyzed to investigate the hydrologic properties at the basin scale, including the speed with which vegetation and soil moisture respond to precipitation anomalies. We also looked at the basin-specific normalized radar surface-backscattering cross-sections from NASA's QuikSCAT Scatterometer, to obtain information on the vegetation regimes of the various Malagasy basins. Finally, we correlated the basin response to the precipitation forcing, and compared the amplitude and time lag of the correlations across watersheds with high elevation headwaters versus those with low elevation headwaters in the aim of evaluating the drought-response hypothesis of Wilmé et al. (2006). Our results indicate that the vegetation water content time series exhibit several features that are consistent with those of the majority of the bioclimatic zones of the island. Specifically, although the speed of the response of the vegetation water content varies significantly among the different basins, it is inter-annually consistent for each watershed, while the soil–moisture time series are less consistent than the vegetation water content time series. This study is a first step in the quantification of the hydrologic properties derived from microwave remote sensing, and which could potentially shed new light on the different intra-annual responses of watersheds to precipitation anomalies. Furthermore, this analysis offers important insights into the hydro-geomorphologic drivers associated with biodiversity patterns in Madagascar, contributing to a better understanding of the mechanisms that determine biotic diversification across the various bioclimatic regions of the island.