RESUMEN
There has been an increase in tile drained area across the US Midwest and other regions worldwide due to agricultural expansion, intensification, and climate variability. Despite this growth, spatially explicit tile drainage maps remain scarce, which limits the accuracy of hydrologic modeling and implementation of nutrient reduction strategies. Here, we developed a machine-learning model to provide a Spatially Explicit Estimate of Tile Drainage (SEETileDrain) across the US Midwest in 2017 at a 30-m resolution. This model used 31 satellite-derived and environmental features after removing less important and highly correlated features. It was trained with 60,938 tile and non-tile ground truth points within the Google Earth Engine cloud-computing platform. We also used multiple feature importance metrics and Accumulated Local Effects to interpret the machine learning model. The results show that our model achieved good accuracy, with 96 % of points classified correctly and an F1 score of 0.90. When tile drainage area is aggregated to the county scale, it agreed well (r2 = 0.69) with the reported area from the Ag Census. We found that Land Surface Temperature (LST) along with climate- and soil-related features were the most important factors for classification. The top-ranked feature is the median summer nighttime LST, followed by median summer soil moisture percent. This study demonstrates the potential of applying satellite remote sensing to map spatially explicit agricultural tile drainage across large regions. The results should be useful for land use change monitoring and hydrologic and nutrient models, including those designed to achieve cost-effective agricultural water and nutrient management strategies. The algorithms developed here should also be applicable for other remote sensing mapping applications.
RESUMEN
Understanding agriculturally co-located solar photovoltaic (PV) installation capacity, practices, and preferences is imperative to foster a future where solar power and agriculture co-exist with limited impacts on food production. Crops and PV panels are often co-located as they have similar ideal conditions for maximum yield. The recent boom in solar photovoltaics is displacing a significant amount of cropland. The literature on agriculturally co-located PV array installations lacks important spatiotemporal details that could help inform future array installations and improve associated policies and incentive programs. This study used imagery from the National Agriculture Imagery Program for object-based analysis (within eCognition Developer), and from Landsat 5 TM, 7 ETM+ and 8 OLI for temporal analysis (using LandTrendr) to identify and characterize non-residential ground-mounted PV arrays in California's Central Valley installed between 2008 and 2018. This dataset includes over 210,000 individually identified panels grouped by mount and installation year into 1006 PV arrays (69% are agriculturally co-located). The most common type of mounting system is fixed-axis, and individual co-located systems tend to be small (0.34 MW). There were fewer single-axis tracking arrays, although the average capacity per system is nearly four times higher (1.20 MW). In total, the mapped arrays accounted for 3.6 GW of capacity and generated a cumulative of 32,700 GWh within the Central Valley during the study period. For the 694 identified agriculturally co-located arrays (2.1 GW), significantly sub-optimal installation practices were observed in the spacing and spatial field placement of the arrays. In terms of crop conversion preferences, commodity crops (pastureland) dominated the total cumulative area converted although specialty crops (orchards) also contributed to a large number of solar installations on cropland. These results provide important details of current PV placement practices; understanding these can help to inform future practices and guide future regulations that might promote solar installations while preserving agricultural production.