Estimation of potential denitrification and its spatiotemporal dynamics in seasonally inundated geomorphic units of a large tropical river using satellite data.

Denitrification in large tropical river systems is likely important for nitrogen retention estimates, but is limited by the need for measurements and the ability to scale these estimates to relate seasonal changes to river geomorphology and discharge. Geomorphic units (GUs), that describe the structure of a river system based on their inundation frequency and vegetation cover, may be useful to characterise features that influence denitrification rates. In this study, we tested the hypothesis that measurements of potential denitrification rate (PDR) using denitrification enzyme assays from different GUs could be used to1) relate PDR to soil, vegetation and different land use and land-cover (LULC) types as controlling factors and 2) that these characteristics could be assessed using remote sensing data to model PDR over a large spatial scale (along a 50 km reach) for the Padma River (Bangladesh). Specifically, 245 PDR measurements were made from the four LULC types with in eight GUs during the dry/winter season 2020. Linear regression using a mixed-modelling approach showed that PDR was highly related to vegetation cover and soil moisture across all GUs. Sentinel-2 data were then used to develop relationships between the Normalised Difference Vegetation Index. (NDVI) and vegetation cover and, specifically, between Sentinel-2 band 11 and soil moisture, which also reasonably described PDR rates. We then used this satellite data to estimate reach-scale PDR in post-monsoon, dry/winter and pre-monsoon seasons. The satellite-based model showed that PDR increased in GUs from post-monsoon 2019 to pre-monsoon 2020. The vegetation islands and the bars were the most important GUs for denitrification in all seasons. The satellite-assisted approach developed in this study can be applied to the GUs in large lowland rivers where inundation occurs frequently.