Analysis of spatiotemporal variability of corn yields using empirical orthogonal functions

Authors: KIM, SEONGYUN - UNIVERSITY OF MARYLAND; DAUGHTRY, CRAIG; Russ, Andrew - Andy; PEDRERA-PADILLA, AURA - RESEARCH AND TRAINING INSTITUTE FOR AGRICULTURAL AND FISHERIES OF ANDALUSIA, IFAPA; Pachepsky, Yakov

Submitted to: Journal of Environmetrics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/25/2020
Publication Date: 11/28/2020
Citation: Kim, S., Daughtry, C.S., Russ, A.L., Pedrera-Padilla, A., Pachepsky, Y.A. 2020. Analysis of spatiotemporal variability of corn yields using empirical orthogonal functions. Journal of Environmetrics. 12(12):3339. https://doi.org/10.3390/w12123339.
DOI: https://doi.org/10.3390/w12123339

Interpretive Summary: Crop yields are variable in space and time and decreasing this variability is an important component of the field crop management. Looking for the patterns of spatial variation that remain stable from year to year can help to determine the environmental factors causing variability and focus on mitigating those factors.. Shallow patchy groundwater and preferential subsurface flow pathways may create spatial variation in water supply to root zone and yields. Our objective was to determine the spatial patterns in interannual variation of corn yields in fields with i) uniform applications of manure plus N fertilizer, ii) uniform applications of N fertilizer, and iii) variable rate (precision) applications fertilizer. We found that the pattern that explains more than half of the variability strongly depends on the distance to the subsurface groundwater pathways at the manure and uniform fertilizer application, but not at the precision farming application field. Eliminating the effect of the subsurface patchy groundwater on the yields led to the savings of about 25% of fertilizer. Accounting for the effect of shallow patchy groundwater and subsurface flow pathways on the spatial variability of yields may be a resource-saving practice in the precision farming field management.

Technical Abstract: Crop yields often show significant patterns of temporal and spatial variability. Empirical Orthogonal Function (EOF) analysis have been used to characterize spatial and temporal patterns of variation in many disciplines. In this paper, we used Empirical Orthogonal Function (EOF) to analyze the spatial and temporal patterns of corn (Zea mays L.) yields of three small (3.4 to 4.1 ha) hydrologically-bounded fields with subsurface preferential lateral flow pathways and different nutrient management practices. One field received uniform applications of manure and N fertilizer; the second field received uniform applications of N fertilizer; and the third field received variable rate (precision) applications of N fertilizer. Six years of grain yield monitoring data were analyzed. Spatial patterns of subsurface flow pathways (EOF1) accounted for 52 to 56% of the interannual variability of corn yields. The second (EOF2) and third (EOF3) spatial patterns which explained only 17 to 20% and 10 to 13%, respectively, were not related to subsurface topography. Clear distinct semivariograms which measured the spatial variability of EOF1 were found for fields with uniform applications of N fertilizers and/or manure. The precision applications of N fertilizer minimized corn yield variability associated with subsurface preferential flow patterns. Investigating spatial variable of yield variability for this type of agricultural control continues to show a promising research perspective.