Exploration of sensor comparability: a case study of composite MODIS Aqua and Terra data

Abstract

Comparability across sensors is a common goal of remotely sensed data products; however, sensor calibration and atmospheric conditions often limit implementation. Recently, a slight degradation in the blue band sensor on the Moderate Resolution Imaging Spectroradiometer (MODIS) Terra satellite was reported (Wang et al. 2012), and it was suggested at the 2011 fall NASA Carbon Cycle and Ecosystems/Biodiversity and Ecological Forecasting Team meeting that users switch to using MODIS Aqua data until the corrected collection 6 Terra data are released. Following this recommendation, a switch from Terra to Aqua data inputs used in the Tsetse Ecological Distribution (TED) model was implemented; however, significant variability between model outputs for Kenya was found. This letter considers the possible sources of incongruity between the MODIS Aqua and Terra sensors that might result in the observed differences between model outputs, including sensor degradation, temporal overlap, daily acquisition and data quality. While for data of Kenya it appears that differences in daily acquisition times and data quality result in differences between composite products, ultimately the selection of what data product and sensor to use should be based on reported data quality and biophysical conditions of the region under study.