A Case Study for the Okavango Delta Wetland: Using Modis Time Series Observations to Ascertain Flooding Regime Dynamics for Improved Methane Emission Modeling Inputs

Remote sensing plays a significant role in the estimation of methane and monitoring wetlands flooding duration and extent [1]. Methane is the second largest contributor to the present day greenhouse gas effect [3]. However there are considerable uncertainties in the magnitude of this biogenic emission [2]. For instance, the amount of methane emissions from the Okavango Delta is not known, which makes it difficult for Botswana to present a comprehensive report to the UNFCC on greenhouse emissions budgets. Therefore, this study presents a method for mapping temporal inundation patterns for improved methane emission estimates in the Okavango Delta using MODIS data. It is the first time that MODIS data is used for this purpose. Previous researches in the Delta relied on Landsat data that has a fine resolution but less temporal coverage as compared to MODIS. Even though MODIS data has a courser resolution (250-1000meter), it is better suited to map intra-annual and inter-seasonal inundation variations especially since the imagery is acquired near to daily and only one image scene covers the whole Delta. The data is freely available online, but without the necessary pre-processing tools it is not possible to make correct scientific conclusions from the time series data, since the standard MODIS compositing algorithm does not preclude ‘noise’ effects related to cloud, cloud shadow and BRDF. We derived a contextual algorithm using local near infrared minima over one year and combined this with an index function between the maximum chlorophyll active vegetation stage in the wetland, using the NDVI band, and low NIR metrics in the maximum flooded season. The resulting inundation length and wetland morphology map was used to determine areas within the delta where field measurements on methane fluxes will be used set up. As a result we attained methane emissions results for each flooding regime class, based on intra-seasonal dynamics of water inundation levels and wetland vegetation morphology.