Determining the number of measurements required to estimate crop residue cover by different methods
Citation
Laamrani, A., Joosse, P., Feisthauer, N. (2017). Determining the number of measurements required to estimate crop residue cover by different methods. Journal of Soil and Water Conservation, [online] 72(5), 471-479. http://dx.doi.org/10.2489/jswc.72.5.471
Plain language summary
Crop residues are materials left in an agricultural field after the crop has been harvested. These residues include corn stalks, corn cobs, wheat straw and other leftovers from soybeans. They play an important role in reducing potential soil erosion and in increasing soil organic matter content of agricultural soils. Crop residue management is a practice of great importance in southwestern Ontario, where soil management practices affect Great Lakes water quality. The use of remote sensing data to measure and monitor crop residue can be fast and efficient. However, prior to using remote sensing-based data, this work aimed to determine the optimal number of field measurements required to accurately estimate residue levels from remote sensing techniques. We first compared the residue estimates derived from digital photographs with those derived from the standard on-the-ground line-transect method, using 18 agricultural fields located in southern Ontario. We then retrospectively compared the minimum number of line-transects or photographs required to estimate crop residue cover per residue level. The results showed (i) a good correlation between percentage crop residue estimates derived from photographs and those derived from line-transects; (ii) the minimum number of samples required to estimate crop residue cover are dependent on the residue level; (iii) a greater variability in the results for soybeans than for corn; and, (iv) medium-level soybean residue has the highest variability. Overall, the results demonstrated that: the photograph method to estimate residue is a suitable alternative to the line-transect method, which is more time-consuming and labour-intensive; and, two to 54 samples (photos or transects) per field were sufficient to accurately estimate the percentage of residue cover. These results are important as they provide guidance for the use of digital photograph capture methods for remote sensing calibration and validation, or for emerging hand-held mobile device applications for measuring crop residue.
Abstract
Crop residue left after harvest plays an important role in controlling against soil erosion and in increasing soil organic matter content of agricultural soils. Crop residue management is a practice of great importance in southwestern Ontario, where soil management practices have an effect on Great Lakes water quality. The use of remote sensing data to measure and monitor crop residue can be fast and efficient. However, remote sensing-based studies need calibration and validation using field observations. The objective of this study was to determine the optimal number of ground-truthing field measurements (i.e., digital photographs) required to estimate residue levels. To do so, we compared the residue estimates derived from digital photographs with those derived from the standard line-transect method. Residue was measured from 18 fields located in southern Ontario, and data collected included percentage of crop residue using line-transect and photographic grid methods. Results were analyzed using linear regression, correlation tests, ANOVA, and means tests. Analyses were also conducted to retrospectively determine the minimum number of line transects or digital photos required to estimate crop residue cover at specified levels of power. Results showed that (1) percentage of crop residue estimates derived from using digital photographs were strongly correlated (r = 0.91, p ≤ 0.001, R2 = 0.83, and n = 90) to those derived from using line transects; (2) counting 50 to 100 points per digital photograph was sufficient to accurately estimate the percentage of residue cover; and (3) there was greater variability in the results for soybean (Glycine max [L.] Merr.) than for corn (Zea mays L.), with the highest variability for medium-level soybean residue. Overall, the digital photograph method to estimate percentage of residue was found to be a suitable alternative to the line-transect method, which is more time consuming and labor intensive. Determining the optimal numbers of measurements to estimate crop residue cover is important to those wishing to use digital photo capture methods to record, archive, and measure residue for remote sensing calibration and validation or for handheld mobile device applications.