Characterizing floods in the poorly gauged wetlands of the Tana River Delta, Kenya, using a water balance model and satellite data

Wetlands, such as those of the Tana River Delta in Kenya, are vital but threatened ecosystems. The flooding characteristics of wetlands largely determine their physical, chemical and biological properties, so their quantification is crucial for wetland management. This quantification can be achieved through hydrological modelling. In addition, the analysis of satellite imagery provides essential hydrological data to monitor floods in poorly gauged zones. The objective of this study was to quantify the main water fluxes and flooding characteristics (extent, duration and number of floods) in the poorly gauged Tana River Delta in East Africa during 2002–2011. To do so, we constructed a lumped hydrological model (the Tana Inundation Model, TIM) that was calibrated and validated with MODIS data. Further analysis of the MYD09A1 500 m composite product provided a map of the empirical probability of flooded state. In non-extreme years and for the current topology of the delta, the flood extent exceeded 300 km2. Floods over 200 km2 occurred on average once a year, with a mean duration of 18 days. River discharge from the upper basin counted for over 95% of the total water inflow. The results are discussed in the light of possible improvements of the models and wetland management issues. This study provides the first known quantification of spatial and temporal flooding characteristics in the Tana River Delta. As such, it is essential for the water and natural resource management of the Tana River basin. The water balance approach was pertinent to the study of this system, for which information on its internal properties and processes is limited. The methodology, a combination of hydrological modelling and flood mapping using MODIS products, should be applicable to other areas, including those for which data are scarce and cloud cover may be high, and where a medium spatial resolution is required.