Abstract: Globally, more people are impacted by extreme hydrologic events such as flooding than all other types of natural disasters combined, and the effects can be devastating. Two examples are the 2012 and 2022 floods along the Niger and Benue Rivers within the Lower Niger River Basin (LNRB) in Nigeria. Flooding within the LNRB typically occurs annually during the rainy season, however, the 2012 and 2022 flood events were of similar magnitude, had catastrophic socioenvironmental impacts, and occurred one decade apart. Limited historical gage data along the Niger and Benue Rivers precludes traditional flood frequency analysis in the LNRB. Hence, this study seeks to utilize globally available observations from satellite remote sensing to compute flood depths using the Floodwater Depth Estimation Tool (FwDETv2.1 version) implemented in Google Earth Engine. Other hydrological and physiographic characteristics of LNRB in 2012 and 2022 are also evaluated using remote sensing observations. Since the FwDET requires only globally available input data (flood inundation map and Digital Elevation Model) which favors data-sparse regions such as Nigeria, the potential for the FwDET tool to automatically quantify flood water depths, an important variable in flood frequency estimation and damage assessment, can be analyzed even when historical observations are lacking. The utility of the FwDETv2.1 for flood management and mitigation studies along global rivers with limited historical data is discussed. ReferenceIdowu, Dorcas, and Wendy Zhou. "Performance evaluation of a potential component of an early flood warning system—A case study of the 2012 flood, Lower Niger River Basin, Nigeria." Remote Sensing 11.17 (2019): 1970.Brakenridge, G. R., Kettner, A. J., Paris, S., Cohen, S., Nghiem, S. V. , River and Reservoir Watch Version 4.5, Satellite-based river discharge and reservoir area measurements, DFO Flood Observatory, University of Colorado, USA. http://floodobservatory.colorado.edu/ SiteDisplays/ 20.htm (Accessed 6 December 2023).Cohen, S.; Peter, B.G.; Haag, A.; Munasinghe, D.; Moragoda, N.; Narayanan, A.; May, S. Sensitivity of Remote Sensing Floodwater Depth Calculation to Boundary Filtering and Digital Elevation Model Selections. Remote Sens. 2022, 14, 5313. https://doi.org/10.3390/rs14215313.B. G. Peter, S. Cohen, R. Lucey, D. Munasinghe, A. Raney and G. R. Brakenridge, "Google Earth Engine Implementation of the Floodwater Depth Estimation Tool (FwDET-GEE) for Rapid and Large Scale Flood Analysis," in IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1-5, 2022, Art no. 1501005, doi: 10.1109/LGRS.2020.3031190.Brakenridge, G. Robert, Son V. Nghiem, and Zsofia Kugler. "Passive microwave radiometry at different frequency bands for river discharge retrievals." Earth and Space Science 10.8 (2023): e2023EA002859.Idowu, Dorcas. Assessing the Utilization of Remote Sensing and GIS Techniques for Flood Studies and Land Use/Land Cover Analysis Through Case Studies in Nigeria and the USA. Diss. Colorado School of Mines, 2021.Idowu, Dorcas, and Wendy Zhou. "Global Megacities and Frequent Floods: Correlation between Urban Expansion Patterns and Urban Flood Hazards." Sustainability 15.3 (2023): 2514.

Globally available, georeferenced data from earth observing satellites and coupled earth systems models provide new opportunities to use limited field observations to infer trends in environmental processes across unsampled locations and over time. Building statistical models that are intended generalize earth surface processes across large spatial scales represents a new frontier in supervised statistical inference. For example, care must be taken to collect unbiased samples given the complexity of the earth system: surface features can appear vastly different from the perspective of multispectral and SAR imagery in different atmospheric/landscape contexts. Environmental processes occur at variable spatial/temporal scales, and sampling resolution can drastically alter the appearance of patterns, and lead to spatial and temporal autocorrelation which can bias model weights and/or parameter estimates. Multicollinearity in multivariate datasets, which is also scale dependent, can inflate variance in parameter/weight estimates. All these in tandem can undermine the robustness of models in predicting in out-of-sample contexts. To overcome this, the GRRIEn (Generalizable, Reproducible, Robust, and Interpretable Environmental) analysis framework is introduced as a standard method of training and validating supervised data-driven models at large spatial scales. The method is explained, and demonstrated with a case study detecting surface water at CONUS scale using SAR and multispectral imagery.