Recipient Organization
UNIVERSITY OF KENTUCKY
500 S LIMESTONE 109 KINKEAD HALL
LEXINGTON,KY 40526-0001
Performing Department
(N/A)
Non Technical Summary
The objective of this sabbatical project is to evaluate the impact of erosion on crop health. The PI postulates that farmers are more willing and interested in adopting erosion control measures and sustainable land management practices when they can quantify the impacts oferosion on crop growth.I propose to spend a sabbatical year developing remote sensing techniques (RGB, multispectral, hyperspectral and thermal imaging) to evaluate crop health on dissected landscapes of the Quapaw Tribal Lands in northeastern Oklahoma. The landcover in the region is predominantly forested, however about 25% of the land, including the reclaimed mine areas, is used for forage and crop production on the terraces, valleys, and floodplains. Very little soil resource data is available on Tribal Lands, however recently, Winzeler et al. (2023) developed a raster-based soil map of the Quapaw Tribal Lands at a 10 m resolution. This information, coupled with a recently developed digital soil map of soil properties at the surface (0-15 cm) and near surface (15-30 cm) of the 22,880-ha reservationmakes this an ideal location to implement this study. As a contrast to these dissected landscapes in the Quapaw Tribal Lands, Co-Host Dr. Daniels manages over 40 edge-of-field water quality monitored watersheds across the Arkansas Discovery Farm network, of which many are located on the same physiographic region as the Quapaw Tribal Lands. Many of these monitored fields are on highly productive farmland with very little erosion making them ideal locations for comparison of erosion impacts on crop health.
Animal Health Component
100%
Research Effort Categories
Basic
(N/A)
Applied
100%
Developmental
(N/A)
Goals / Objectives
APPLICANT CAPABILITIES ENHANCEMENTThe PI will utilize this 1-year sabbatical to hone remote sensing techniques, broaden experienceswith the utilization of vegetative indices, and develop ground reference methods regarding erodedlandscapes and crop health. Spending a year with experts on well characterized study sites willprovide him with the skills needed to adapt these methods to his current and ongoing Blue WaterFarms program in western Kentucky where erosion is actively being observed (Figure 1 and 2).FUTURE RESEARCH GOALS POST-SABBATICALFuture goals post-sabbatical include incorporating these remote sensing techniques into the BlueWater Farms network workflow on ten of our actively eroding, monitored watersheds. Thesetechniques will not only help us quantify the infield erosion and the crop health changes associatedwith erosion that have occurred across these monitored watersheds, but also help us extrapolatethese results from intensively monitored areas to the larger field landscape scale. Theseobservations will be shared with participating growers, commodity boards, and countyCooperative Extension Service agents with the expectation that farm managers will recognizeerosion as having a negative impact on crop health as well as a resource loss, leading to theadoption of conservation BMPs and/or management strategies that will protect the long-termsustainability of their farms, the agricultural community, and the environment. Preliminary datafrom this future work will be used to develop a USDA AFRI Integrated Research/Extensionproposal to expand our existing Blue Water Farms network into forage operations in westernKentucky.
Project Methods
I plan to evaluate the impacts of erosion on crop health through a growing season on privatelyowned farms with eroded row crop fields and forage fields utilizing remote sensing and vegetativeindices. Small (1-6 ha) watersheds with significant erosion and limited erosion will be selectedthrough discussions and site evaluations with Dr. Owens (USDA ARS) who has access to workingfarms with dissected terraces on Quapaw Tribal Lands, and Dr. Daniels (UA) who has access toworking farms with limited erosion participating in the Arkansas Discover Farms EoF waterquality monitoring network. Three each of representative eroded row cropped watersheds anderoded forage watersheds will be selected for evaluation. Erosion will be characterized via sUASLiDAR, Order 1 soil surveys, A horizon thickness investigation and carbon measurements of thesurface horizon. One limited erosion watershed under row crop and forage watershed from theArkansas Discovery Farms network will be evaluated for comparison for a total of 8 studywatersheds. All watersheds will be imaged through the growing season (approximately every 2weeks post-planting or more frequently under environmentally stressing conditions, i.e. earlyplanting, intensive rainfall, drought) via remote sensing satellite platform (i.e. Sentinal-2) and asUAS platform (i.e. X-Wing UAV with Micasense Altum multispectral camera)) to developvegetative indices indicative of crop health. Data will be processed with image processingsoftware (i.e. Pix4D). Soil moisture samples will be collected at representative landscape positionswithin each monitored watershed at the time of each sUAS data collection mission where a thermalcamera is involved (i.e. Altum PT).Vegetative indices (VIs) obtained from remote sensing of plant canopies are an effective way toconduct quantitative and qualitative evaluations of vegetation cover, plant vigor, and growthdynamics (Xue and Su, 2017). There are two common platforms utilized to collect this data,satellites and sUAS. Although satellite platforms have free access to visible and multispectral data,such as Landsat 7-8, there are limitations for agricultural applications which include a coarseresolution (30 m2 for Landsat and 500 m2 for MODIS) and a lengthy revisitation time (up to 2weeks or more depending on cloud cover). Sentinal-2 is another satellite remote sensing platformthat has improved resolution (up to 10 m) and a 5-day revisitation time, however this time delaybetween data sampling can prove challenging for nutrient and water related agriculturalmanagement decisions. Alternative to satellites, RGB, multispectral, and hyperspectral sensorshave been developed to fit onto sUAS and proven capable of providing similar information asremotely sensed data from satellites. Although sUAS imagery for crop health can provide morecontrol over image collection, however it is more expensive, requires a licensed operator, fairweather, and not all areas are sUAS accessible.Vegetative indices will be compared to ground based crop growth indicators (i.e. hay biomassproduction, crop growth stage days post-planting) and soil moisture throughout the growing seasonin eroded and non-eroded areas within each study watershed along transects on each representativelandscape positions (i.e. summit, backslope, footslope) perpendicular to erosion channel on theQuapaw Tribal Lands study watersheds and representative zones on the non-eroded watershedswithin the Arkansas Discover Farms network. There are a number of existing vegetative indicesthat can be calculated using multispectral and RGB sensors including Normalized DifferenceVegetation Index, NDVI (Rouse et al., 1974), Red Edge Normalized Difference Vegetation Index,NDVIred edge (Gitelson and Merzlyak 1994), the Enhanced Vegetation Index 2, EVI2 (Huete, 2002),the Red Edge Simple Ratio, SRred edge, the Green and Red Edge Chlorophyll Indices, CIgreen andCIred edge (Gitelson and Merzlyak, 2003), the MERIS Terrestrial Chlorophyll Index, MTCI (Dashand Curran, 2004), and the Core Red Edge Triangular Vegetation Index (RTVIcore).