No abstract

The Integrated Multisatellite Retrievals for Global Precipitation Measurement (GPM) (IMERG) Level 3 estimates rainfall from passive microwave sensors onboard satellites that are associated with several uncertainty sources such as sensor calibration, retrieval errors, and orographic effects. This study aims to provide a comprehensive investigation of multiple machine learning (ML) techniques (Random Forest, and Neural Networks), to stochastically generate an error-corrected improved IMERG precipitation product at a daily time scale and 0.1°-degree spatial resolution over the Brahmaputra river basin. In this study, we used the operational IMERG-Late Run version 06 product along with several meteorological and land surface parameters (elevation, soil type, land type, soil moisture, and daily maximum and minimum temperature) to produce an improved precipitation product in the Brahmaputra basin. We trained, tested, and optimized ML algorithms using 4 years (from 2015 through 2019) of reference rainfall data derived from the rain gauge. The ML generated precipitation product exhibited improved systematic and random error statistics for the study area, which is a strong indication for using the proposed algorithms in retrieving precipitation across the globe. We conclude that the proposed ML-based ensemble framework has the potential to quantify and correct the error sources for improving and promoting the use of satellite-based precipitation estimates for water resources applications.

Despite the potential of remote sensing for monitoring reservoir operation, few studies have investigated the extent to which reservoir releases can be inferred across different spatial and temporal scales. Through evaluating 21 reservoirs in the highly regulated Greater Mekong region, remote sensing imagery was found to be useful in estimating daily storage volumes for within‐year and over‐year reservoirs (correlation coefficients [CC] ≥ 0.9, normalized root mean squared error [NRMSE] ≤ 31%), but not for run‐of‐river reservoirs (CC < 0.4, 40% ≤ NRMSE ≤ 270%). Given a large gap in the number of reservoirs between global and local databases, the proposed framework can improve representation of existing reservoirs in the global reservoir database and thus human impacts in hydrological models. Adopting an Integrated Reservoir Operation Scheme within a multi‐basin model was found to overcome the limitations of remote sensing and improve streamflow prediction at ungauged cascade reservoir systems where previous modeling approaches were unsuccessful. As a result, daily regulated streamflow was predicted competently across all types of reservoirs (median values of CC = 0.65, NRMSE = 8%, and Kling‐Gupta efficiency [KGE] = 0.55) and downstream hydrological stations (median values of CC = 0.94, NRMSE = 8%, and KGE = 0.81). The findings are valuable for helping to understand the impacts of reservoirs and dams on streamflow and for developing more useful adaptation measures to extreme events in data sparse river basins.

Bangladesh is home to a network of hundreds of rivers and the world's largest river delta, the Ganges Delta. Historically, the nation has been water rich. But that is changing owing to declining rainfall, more-intensive irrigation and heavier use of water upstream. Contamination from arsenic and sewage is also on the rise.To feed our future planet, it is crucial that water is used more sustainably in agricultural regions such as Bangladesh. Other agricultural hotspots face similar water stresses, including the Western and Central United States, northern India and Brazil 1,2 , where falling water tables punish farmers and grab headlines.Bangladesh has taken some steps to address the problem. In 2018, its Ministry of Planning published the Bangladesh Delta Plan 2100 (BDP; see go.nature.com/3s26anc). This outlines a long-term strategy for the country's sustainable and resilient socio-economic development in a changing climate. Water security is a key part of this plan. Although the BDP rightly identifies the main issues facing the nation's water, it is vague on effective actions. These will require heavy investments and more supporting research.Intensive irrigation and climate change are depleting groundwater reserves in this fast-developing nation. To improve its water security, researchers need more information on water use, quality, flows and forecasts.

A computationally efficient early warning technique is developed for forecasting flash floods during the pre-monsoon season that are associated with a complex topography and transboundary runoff in northeastern Bangladesh. Locally conditioned topographic and hydrometeorological observations are key forcings to the modeling system that simulate the hydrology and hydraulic processes. The hydrologic model is calibrated and validated using satellite-based observations to estimate the correct amount of transboundary and mountainous inflow into the flash flood-prone plains. Inflow is then forecasted using precipitation forecast from a global numerical weather prediction (NWP) system called the Global Forecasting System (GFS). The forecasted inflows serve as the upstream boundary conditions for the hydrodynamic model to forecast the water stage and inundation downstream in the floodplains. A real-time in-situ data-based error correction methodology is applied to maintain the skill of the system. The simulation grid size and time-step of the hydrodynamic model are also optimized for computational efficiency. Historical performance of the framework revealed at least 60% accuracy at 5-day lead-time in delineating flood inundation when compared against Sentinel-1 synthetic aperture radar (SAR) imagery. The study suggests that higher resolution topographic information and dynamically downscaled meteorological observations can lead to significant improvement in flash flood forecasting skills.