Frost prediction using machine learning and deep neural network models.

This study describes accurate, computationally efficient models that can be implemented for practical use in predicting frost events for point-scale agricultural applications. Frost damage in agriculture is a costly burden to farmers and global food security alike. Timely prediction of frost events is important to reduce the cost of agricultural frost damage and traditional numerical weather forecasts are often inaccurate at the field-scale in complex terrain. In this paper, we developed machine learning (ML) algorithms for the prediction of such frost events near Alcalde, NM at the point-scale. ML algorithms investigated include deep neural network, convolution neural networks, and random forest models at lead-times of 6-48 h. Our results show promising accuracy (6-h prediction RMSE = 1.53-1.72°C) for use in frost and minimum temperature prediction applications. Seasonal differences in model predictions resulted in a slight negative bias during Spring and Summer months and a positive bias in Fall and Winter months. Additionally, we tested the model transferability by continuing training and testing using data from sensors at a nearby farm. We calculated the feature importance of the random forest models and were able to determine which parameters provided the models with the most useful information for predictions. We determined that soil temperature is a key parameter in longer term predictions (>24 h), while other temperature related parameters provide the majority of information for shorter term predictions. The model error compared favorable to previous ML based frost studies and outperformed the physically based High Resolution Rapid Refresh forecasting system making our ML-models attractive for deployment toward real-time monitoring of frost events and damage at commercial farming operations.

Predicting Chronic Climate-Driven Disturbances and Their Mitigation.

Society increasingly demands the stable provision of ecosystem resources to support our population. Resource risks from climate-driven disturbances, including drought, heat, insect outbreaks, and wildfire, are growing as a chronic state of disequilibrium results from increasing temperatures and a greater frequency of extreme events. This confluence of increased demand and risk may soon reach critical thresholds. We explain here why extreme chronic disequilibrium of ecosystem function is likely to increase dramatically across the globe, creating no-analog conditions that challenge adaptation. We also present novel mechanistic theory that combines models for disturbance mortality and metabolic scaling to link size-dependent plant mortality to changes in ecosystem stocks and fluxes. Efforts must anticipate and model chronic ecosystem disequilibrium to properly prepare for resilience planning.

GRACE satellite observations reveal the severity of recent water over-consumption in the United States.

Changes in the climate and population growth will critically impact the future supply and demand of water, leading to large uncertainties for sustainable resource management. In the absence of on-the-ground measurements to provide spatially continuous, high-resolution information on water supplies, satellite observations can provide essential insight. Here, we develop a technique using observations from the Gravity Recovery and Climate Experiment (GRACE) satellite to evaluate the sustainability of surface water and groundwater use over the continental United States. We determine the annual total water availability for 2003-2015 using the annual variability in GRACE-derived total water storage for 18 major watersheds. The long-term sustainable water quantity available to humans is calculated by subtracting an annual estimate of the water needed to maintain local ecosystems, and the resulting water volumes are compared to reported consumptive water use to determine a sustainability fraction. We find over-consumption is highest in the southwest US, where increasing stress trends were observed in all five basins and annual consumptive use exceeded 100% availability twice in the Lower Colorado basin during 2003-2015. By providing a coarse-scale evaluation of sustainable water use from satellite and ground observations, the established framework serves as a blueprint for future large-scale water resource monitoring.