Bryan C. Runck

Improving evapotranspiration (ET) estimates is critical for developing accurate, region-specific irrigation scheduling tools. Our objective is to improve actual evapotranspiration (ETₐ) estimation for corn and soybean in Minnesota using satellite-based data from OpenET, Climate Engine and the USDA cropland data layer. Monthly ETₐ from OpenET was combined with daily reference evapotranspiration (ET₀) from Climate Engine to calculate region-specific crop coefficients (Kc) for irrigated corn and soybean. Seasonal ETₐ for corn and soybean in southern and central counties ranged from 500 to 590 mm (2018–2023). Daily corn Kc values across 25 Minnesota counties ranged from 0.39–0.56 at initial stage Days after planting (DAP 0–25), peaked at 0.8–0.92 during mid-season (DAP 75–125), and declined to 0.46–0.62 by late season (DAP 150–175). Soybean Kc values ranged from 0.38–0.58 in the initial stage (DAP 0–25), increased through canopy development to a mid-season peak of 0.78–0.91 (DAP 75–125), and declined to 0.43–0.56 by late season (DAP 150–175). Ground-based validation using Eddy covariance sensors at the V8, I10, and I18 sites showed strong agreement with OpenET-derived ETₐ, with R² = 0.64 (2021), R² = 0.70 (2022) and R² = 0.66 (2023) for V8, R² values of 0.60, 0.31, and 0.55 for I10 (2020–2022), and R² values ranging from 0.53 to 0.66 for I18 South (2018–2024) at the daily time step. By combining satellite-based ETₐ with ground validation, this study generates localized Kc values that support region-specific irrigation scheduling and improved water-use efficiency in Minnesota.

Sky islands are high-elevation ecosystems surrounded by lowland habitats that create isolated environments with distinct climatic conditions. These factors have driven the evolution of many endemic species, separated from their larger, contiguous populations. An Individual-Based Model (IBM) was used to simulate population dynamics by modeling the behaviors and interactions of Tamiasciurus hudsonicus grahamensis (Mount Graham Red Squirrel) a subspecies of the American red squirrel (Tamiasciurus hudsonicus) that is endemic to the Pinaleno Mountains in southeastern Arizona. This approach can help predict future population trends based on historical species data leading to better conservation decisions. Using species-specific ecological preferences - including temperature, precipitation, and vegetation indices (NDVI) - an IBM was developed to simulate population dynamics and spatial distribution projections through 2100. Climate change projections, based on the best- and worst-case scenarios outlined in the 2014 National Climate Assessment, were incorporated to assess potential future population trends under changing environmental conditions. The population faces a 45-62% probability of extinction by 2100, with a significant risk of extinction within the next 50 years. A translocation experiment was conducted to evaluate the viability of relocating individuals to the Chiricahua Mountains, another sky island with a larger habitable area. However, the risk of extinction remains even higher (87-89%) due to environmental disturbances affecting both the Chiricahua and Pinaleno regions. This highlights the challenges of conservation efforts in the face of climate change and emphasizes the need for targeted management strategies to preserve this critically endangered subspecies.

Climate change is altering crop phenology, temperature regimes, and water availability, posing challenges for effective agricultural water management. We focus on corn (VT-R1), soybean (R3-R4), and potato (tuber initiation-bulking) to evaluate whether precipitation has historically met evapotranspiration demands during these key sensitive growth stages in Central Minnesota. We analyzed accumulated growing degree days (AGDD, 1951–2024) and precipitation using planting windows of April 15–25 for potatoes and May 1–10 for corn and soybean. Field observations from 2023 and 2024 at Becker and Saint Paul confirm that recent years have more AGDD than historical norms. We then compare empirical field measurements with historical AGDD and precipitation data to estimate crop ET demands during the growth stages most sensitive to water stress. We found that rainfall rarely matched ET demands during these stages from 1951–2024, as median rainfall frequently fell below growth stage-specific ET thresholds. The high interannual and spatial variability meant that precipitation was not consistently aligned with peak crop demand. We also find no trend of increasing or decreasing water stress over this time period. These findings highlight that corn, soybean and potato grown in Central Minnesota generally have not met the crop-water demand during the most-sensitive growth stages without irrigation. Crop production in the Central Sand Plains of Minnesota has not gotten riskier since 1951; it is just as risky as it has been historically requiring yearly growth stage-specific irrigation scheduling aligned with crop growth stages and precipitation to sustain yields and build climate-resilient agricultural systems.

Evapotranspiration (ET) plays a critical role in the land-atmosphere interactions, yet its accurate quantification across various spatiotemporal scales remains a challenge. In situ measurement approaches, like eddy covariance (EC) or weather station-based ET estimation, allow for measuring ET at a single location. Agricultural uses of ET require estimates for each field over broad areas, making it infeasible to deploy sensing systems at each location. This study integrates tree-based and knowledge-guided machine learning (ML) techniques with multispectral remote sensing data, griddled meteorology and EC data to upscale ET across the Midwest United States. We compare four tree-based models - Random Forest, CatBoost, XGBoost, LightGBM - and a simple feed-forward artificial neural network in combination with features engineered using knowledge-guided ML principles. Models were trained and tested on EC towers located in the Midwest of the United States using k-fold cross validation with k=5 and site-year, biome stratified train-test split to avoid data leakage. Results show that LightGBM with knowledge-guided features outperformed other methods with an R2=0.86, MSE=14.99 W m^-2 and MAE = 8.82 W m^-2 according to grouped k-fold validation (k=5). Feature importance analysis shows that knowledge-guided features were most important for predicting evapotranspiration. Using the best performing model, we provide a data product at 500 m spatial and one-day temporal resolution for gridded ET for the period of 2019-2024. Intercomparison between the new gridded product and state-level weather station-based ET estimates show best-in-class correspondence.

Agrometeorological data are essential for understanding production using digital agriculture techniques. However, integrating agrometerological observations from multiple sources remains a challenge. Often, digital agriculture scientists download and clean the same datasets many times. We present a prototype system that simplifies the process of collecting, cleaning, integrating, and aggregating data from meteorological data sources by providing a simplified user interface, database, and application programming interface. The prototype provides a standard interface for querying multiple geospatial formats (raster and vector) and integrates observation networks including the National Oceanic and Atmospheric Administration Global Historical Climatology Network (NOAA GHCN), NOAA NClim‐Grid (NOAA's Gridded Climate Normals), and Ameriflux BASE. The system automatically checks and updates data, saving storage space and processing time, and allows users to summarize data spatially and temporally. Provided as open source code and browser‐based user interface, the application and integration system can be run across Windows, Linux, and Mac environments to support broader use of multi‐source agrometeorology data.

Plant cover and biochemical composition are essential parameters for evaluating cover crop management. Destructive sampling or estimates with aerial imagery require substantial labor, time, expertise, or instrumentation cost. Using low‐cost consumer and mobile phone cameras to estimate plant canopy coverage and biochemical composition could broaden the use of high‐throughput technologies in research and crop management. Here, we estimated canopy development, tissue nitrogen, and biomass of medium red clover (Trifolium pratense L.), a perennial forage legume and common cover crop, using red‐green‐blue (RGB) indices collected with standard settings in non‐standardized field conditions. Pixels were classified as plant or background using combinations of four RGB indices with both unsupervised machine learning and preset thresholds. The excess green minus red (ExGR) index with a preset threshold of zero was the best index and threshold combination. It correctly identified pixels as plant or background 86.25% of the time. This combination also provided accurate estimates of crop growth and quality: Canopy coverage correlated with red clover biomass (R² = 0.554, root mean square error [RMSE] = 219.29 kg ha⁻¹), and ExGR index values of vegetation pixels were highly correlated with clover nitrogen content (R² = 0.573, RMSE = 3.5 g kg⁻¹) and carbon:nitrogen ratio (R² = 0.574, RMSE = 1.29 g g⁻¹). Data collection were simple to implement and stable across imaging conditions. Pending testing across different sensors, sites, and crop species, this method contributes to a growing and open set of decision support tools for agricultural research and management.

Agriculture is a designed system with the largest areal footprint of any human activity. In some cases, the designs within agriculture emerged over thousands of years, such as the use of rows for the spatial organization of crops. In other cases, designs were deliberately chosen and implemented over decades, as during the Green Revolution. Currently, much work in the agricultural sciences focuses on evaluating designs that could improve agriculture's sustainability. However, approaches to agricultural system design are diverse and fragmented, relying on individual intuition and discipline-specific methods to meet stakeholders' often semi-incompatible goals. This ad-hoc approach presents the risk that agricultural science will overlook nonobvious designs with large societal benefits. Here, we introduce a state space framework, a common approach from computer science, to address the problem of proposing and evaluating agricultural designs computationally. This approach overcomes limitations of current agricultural system design methods by enabling a general set of computational abstractions to explore and select from a very large agricultural design space, which can then be empirically tested.

Digital agriculture is often heralded as the next major wave of innovation in food and agriculture. Driven by big data, the intention is that digital agriculture will transform the entire research and development pipeline across agricultural value chains throughout the developed and developing worlds. Yet, issues concerning data quality, interoperability, intellectual property ownership and data privacy present considerable challenges to this vision. Digital agriculture platforms, which support data sharing, analysis, interoperability, and public and private sector collaboration, are one approach to address this challenge, but as a new research domain, there is a lack of conceptual clarity around what constitutes a “digital agriculture platform”. Here, we use a “bottom up” and “top down” analysis approach to develop a taxonomy of the digital agriculture landscape. Then, we select a set of digital agriculture platforms for in‐depth analysis across a set of technical and use requirements. While digital agriculture will remain a constantly evolving landscape with varied technologies and application areas, this presents a first attempt to characterize this landscape and establish a common vocabulary for understanding digital agriculture platforms.

Cover cropping is considered a cornerstone practice in sustainable agriculture; however, little attention has been paid to the cover crop production supply chain. In this Perspective, we estimate land use requirements to supply the United States maize production area with cover crop seed, finding that across 18 cover crops, on average 3.8% (median 2.0%) of current production area would be required, with the popular cover crops rye and hairy vetch requiring as much as 4.5% and 11.9%, respectively. The latter land requirement is comparable to the annual amount of maize grain lost to disease in the U.S. We highlight avenues for reducing these high land use costs.