Since 1980, the U.S. has experienced 391 climate disaster events that each caused over $1 billion in losses. Droughts account for 13% of these disasters, with total losses over $360 billion. Accelerated climate change and development in areas vulnerable to climate impacts are escalating these water security risks, requiring urgent shifts in scientific research towards more actionable information for decision-making and climate resilience. A key challenge to developing science-based solutions for these events is creating robust data and tools that provide relevant policy and water management information that are easy to access and interpret, transparent in their development and use, and tailored for stakeholder needs. We present three of our ongoing interdisciplinary research efforts leveraging advancements in land surface modeling (LSM) and Earth observations (EOs) to address water security challenges across North America. First, NASA’s Land Information System team is developing phase three of the North American Land Data Assimilation System (NLDAS-3). NLDAS is a crucial LSM environment supporting water resource management, drought monitoring, and other uses. Building on stakeholder feedback and progress in data assimilation (DA), NLDAS-3 will provide more timely, robust, and accessible information for water resource operations and research. Next, we highlight a collaborative effort between university researchers and water resource managers for long-term planning in the Colorado River Basin. This multi-phase work resulted in future hydrologic scenarios that were incorporated into drought shortage negotiations, a user-centered web tool with interactive analyses of future hydrology scenarios under forest disturbance and climate uncertainties, and ongoing development of a hydrologic monitoring and early warning forecast system. Finally, we detail efforts to disentangle human and climate influences on agricultural drought in the Western U.S. This work integrates LSM, DA, and data-driven analyses to provide metrics for enhanced drought monitoring and resilience, extendable to other regions. Through these examples, we will discuss lessons learned, challenges, and opportunities to inform environmental research seeking similar actionable solutions relevant across sectors and globally.

Accurate characterization of surface meteorological distributions over coastal areas and complex terrain, especially the relationship between temperature and altitude, is essential for simulating snowpack dynamics. This is challenging at spatial resolutions smaller than common gridded meteorological datasets (e.g., resolutions smaller than 5 to 250 km) due to sparse long-term temperature measurements at those resolutions and local factors like cool air pooling and inversions. Near-surface air temperatures (Ta) are often assumed to decrease with elevation at a constant rate of 6.5°C km-1, leading to significant model errors in snow evolution and other key processes. This study evaluated the impact of local dynamical adjustments to downscaled Ta on snow simulations over two coastal mountainous terrains using the Noah-MultiParameterization land surface model. Forcings were derived from remote sensing and reanalysis precipitation products and Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2) atmospheric products at the downscaled 1-km resolution. Hourly lapse rates at each grid cell were calculated by applying linear regression to Ta and elevation from neighbor grids (one grid length in the x or y direction) at the native MERRA-2 resolution and applied to the downscaled 1-km Ta product. We assessed the impact on simulated snow cover and depth across simulations forced with the downscaled Ta (1) without lapse rate correction, (2) corrected with a static lapse rate (6.5°C km-1), and (3) corrected with the dynamic hourly lapse rate. We evaluated model skill improvement with dynamic or static lapse rate correction, and no correction against satellite-derived products. Both lapse rate correction methods led to similar improvements on average, relative to no correction. However, dynamic lapse rate correction showed more pronounced improvements in simulating perennial snowpacks at mid-elevations and in deficit years, indicating the method can better resolve heterogeneous snowpack conditions that are key for water resource management.

Snow reanalyses that combine process-based models and remote sensing observations of snow provide estimates of seasonal snow water equivalent (SWE) evolution that surpass the accuracies of traditional modeling approaches. However, snow reanalyses are only available over smaller subregions, and sometimes use computationally expensive modeling approaches. We investigate whether 1 km-resolution and daily SWE from a popular reanalysis could be learned by connecting only trusted meteorological fields (multidecadal precipitation patterns and daily air temperature) and remotely sensed snow cover using a deep learning model. Relative to point observations of SWE evolution in the western United States, the lightweight deep learning model was able to reproduce the spatial and temporal evolution estimated by the snow reanalysis. Further, we found that the deep learning model could be trained in the western United States and then reused to estimate SWE evolution in the European Alps, demonstrating a high average coefficient of correlation (0.81) and low peak-SWE bias (< 1%) versus point estimates of SWE. SWE from the deep learning model also outperformed SWE estimates from physically based land surface simulations, capturing elevation-driven impacts on SWE spatial heterogeneity and interannual differences in seasonal SWE magnitudes important for water resources, climate regulation, and local ecology. This study demonstrates how deep learning approaches could be used to mine connections between daily SWE evolution, snow cover remote sensing, and limited meteorological information to generate and expand the geographical extent of fine-resolution historical snow estimates in complex terrains.