Search for: All records

Climate models today depend critically on confident initial conditions, a reasonably plausible snapshot of the Earth from which all future predictions emerge. However, given the inherently chaotic nature of our system, this constraint is complicated by sensitivity dependence, where small uncertainties can lead to exponentially diverging outcomes over time. This challenge is particularly salient at global spatial scales and over centennial timescales, where data gaps are not just common but expected. The source of uncertainty is two-fold: (1) sparse, noisy observations from satellites and ground stations, and (2) variability stemming from simplifying approximations within the models themselves. In practice, data assimilation methods are used to reconcile this missing information by conditioning model states on available observations. Our work builds on this idea but operates at the extreme end of sparsity. We propose a conditional data imputation framework that reconstructs full temperature fields from as little as 1% observational coverage. The method leverages a diffusion model guided by a prekriged mask, effectively inferring the full-state fields from minimal data points. We validate our framework over the Southern Great Plains, focusing on afternoon through night (12:00 PM–12:00 AM) temperature fields during the summer months of 2018–2021. Across varying observational densities—from swath data to isolated in situ sensors—our model achieves strong reconstruction accuracy, highlighting its potential to fill in critical data gaps in both historical reanalysis and real-time forecasting pipelines. 

Abstract Dust transported from rangelands of the Southwestern United States (US) to mountain snowpack in the Upper Colorado River Basin during spring (March‐May) forces earlier and faster snowmelt, which creates problems for water resources and agriculture. To better understand the drivers of dust events, we investigated large‐scale meteorology responsible for organizing two Southwest US dust events from two different dominant geographic locations: (a) the Colorado Plateau and (b) the northern Chihuahuan Desert. High‐resolution Weather Research and Forecasting coupled with Chemistry model (WRF‐Chem) simulations with the Air Force Weather Agency dust emission scheme incorporating a MODIS albedo‐based drag‐partition was used to explore land surface‐atmosphere interactions driving two dust events. We identified commonalities in their meteorological setups. The meteorological analyses revealed that Polar and Sub‐tropical jet stream interaction was a common upper‐level meteorological feature before each of the two dust events. When the two jet streams merged, a strong northeast‐directed pressure gradient upstream and over the source areas resulted in strong near‐surface winds, which lifted available dust into the atmosphere. Concurrently, a strong mid‐tropospheric flow developed over the dust source areas, which transported dust to the San Juan Mountains and southern Colorado snowpack. The WRF‐Chem simulations reproduced both dust events, indicating that the simulations represented the dust sources that contributed to dust‐on‐snow events reasonably well. The representativeness of the simulated dust emission and transport in different geographic and meteorological conditions with our use of albedo‐based drag partition provides a basis for additional dust‐on‐snow simulations to assess the hydrologic impact in the Southwest US.