Post-emergence nitrogen (N) fertilizer is typically split applied to irrigated potato (Solanum tuberosum L.) in Minnesota in order to minimize the likelihood of nitrate leaching and to best match N availability to crop demands. Petiole nitrate-nitrogen (NO3-N) concentration is often used as a diagnostic test to determine the rate and timing of split applications, but determining spatial variability within a field using this approach is difficult. Canopy-level spectral measurements, such as hyperspectral and multispectral imagery, have the potential to be a reliable tool for making in-season N management decisions for precision agriculture applications. A two year field study was conducted on a loamy sand soil to evaluate the effects of variety, N treatment, and water stress on: (i) tuber yield, tuber quality, and plant N uptake characteristics; (ii) total N and NO3-N concentrations in petiole, leaflet, and whole leaf tissue samples throughout the season; and (iii) the ability of spectral data, especially from aerial hyperspectral and multispectral imagery, to predict leaf N concentration and tuber yield. The study included two potato varieties (Russet Burbank and Alpine Russet), two irrigation regimes (unstressed and stressed), and five N treatments categorized by three N rates (34 kg N ha-1, 180 kg N ha-1, and 270 kg N ha-1) in which the 270 kg N ha-1 rate had post-emergence N either split applied or applied early in the season. In addition, one of the 270 kg N ha-1 rates with post-emergence N split applied included a soil surfactant application. Reflectance measurements were made using both ground- and aerial-based platforms using several sensors: SPAD-502 chlorophyll meter, Cropscan MSR16R multispectral radiometer, AISA-Eagle hyperspectral camera, and Redlake MS4100 multispectral camera.
Insufficient supplemental water during critical growth stages was found to negatively affect tuber yield, tuber quality, and plant N uptake, but surprisingly had a positive overall effect on frying quality. Because of its high yields, higher proportion of tubers >170g, higher N uptake, and similar processing qualities compared with Russet Burbank (RB), the Alpine Russet (AR) variety is attractive from both an economic and environmental perspective. For all N treatments except the 34 N early treatment, N use efficiency and N uptake efficiency values were less than 50 g g-1 and 50%, respectively. This indicates a relatively low efficiency of N use during 2010 and 2011 attributed to heavy leaching events during these years. Within the same N rates, N uptake efficiency was higher when post-emergence N was split applied, which is likely the result of less N being lost due to leaching. The use of the soil surfactant had a minimal effect on the parameters measured in this experiment.
Tissue samples analyzed for NO3-N were very responsive to N fertilizer applications that occurred within about 7 days of the sampling date, and are therefore a good indicator of current plant N uptake. Alternatively, tissue samples analyzed for total N were more stable over sample dates, and appear to be a better indicator of overall plant N uptake at the time of sampling. In general, ground-based spectral data could predict tuber yield better than any of the N-based tissue sampling procedures; their coefficient of determination (r2) values ranged from 0.40-0.85 on all measurement dates throughout the 2010 and 2011 seasons. Applying the nitrogen sufficiency index (NSI) to plant measurements and spectral indices/models made them mostly insensitive to the effects of variety; this is an effective way to normalize data based on local growing conditions and cultural practices. The best predictor of N stress was determined to be the partial least squares regression model using derivative reflectance as input for its independent variables (r2 of 0.79 for RB and 0.77 for AR). However, the best technique for determining N stress level for variable rate application of N fertilizer using the NSI was determined to be MTCI (MERIS Terrestrial Chlorophyll Index) because the combination of its good relationship with leaf N concentration and high accuracy. The inherent variability of a spectral index should be considered before determining the N sufficiency threshold for determining the rate and timing of post-emergence N fertilizer applications. GRVI (Green Ratio Vegetation Index) normalized by an NSI that used the recommended rate and timing from the research plots as a reference could detect areas of the commercial field that were most unsuitable for supplemental N fertilizer applications on different measurement dates. On 56 and 79 DAE in 2011, most of the commercial field was above the GRVI NSI over-sufficiency threshold level of 120% (the mean pixel values were greater than 127% for both varieties). Because of differences in potato variety, growth stage, sensors, or other local conditions, reference areas are needed in order to make accurate recommendations. The results from this study suggest that diagnostic criteria based on biomass and nutrient concentration (e.g., canopy-level spectral reflectance data) were best suited to determine overall crop N status for determination of in-season N fertilizer recommendations.
University of Minnesota M.S. thesis. November 2012. Major: Soil Science (Land and atmospheric science). Advisors: Carl J. Rosen and David J. Mulla. 1 computer file (PDF); xii, 267 pages appendix p. 246-267.
Nigon, Tyler John.
Aerial imagery and other non-invasive approaches to detect nitrogen and water stress in a potato crop.
Retrieved from the University of Minnesota Digital Conservancy,
Content distributed via the University of Minnesota's Digital Conservancy may be subject to additional license and use restrictions applied by the depositor.