Evaluation and calibration of field techniques to quantify crop cover

Abstract

Reliable estimates of the percentage of the soil surface covered by crop residues and shaded by crop canopies are required for the management of soils to reduce erosion. The importance of ground cover is reflected in its role in erosion prediction models as the cover-management factor. Research has shown that detachment of soil particles by raindrops is the prerequisite to erosion, and that ground cover is the greatest deterrent to this process through raindrop interception and attenuation. There also occurs a critical cover value below which erosion is significant and above which it is insignificant. Most techniques of estimating ground cover are not very precise. Consequently, conservation options derived from predictive models are likely to be erroneous due to inaccurate cover management factors. This research was therefore meant to evaluate and calibrate four different cover measurement techniques (i.e. the photographic method, the sighting-frame. the ceptometer, and the metre-stick method). These techniques were evaluated on maize, bean and maize-bean intercrop cover in agro-c1imatic zone III (semi-humid). All pairs of technique comparisons gave very high correlation coefficients (r) in the range of 0.95 - 0.99. This suggested that any of the four methods could be used for routine ground cover estimation. Although all the techniques gave higher estimates of cover than the actual cover, calibration of the crop cover values against the physical plant models gave r-values of 0.921, 0.909, 0.689, and 0.467 for the camera, sighting frame, metre stick, and ceptometer respectively. The cover XlI correction factors obtained from the regression equations were 0.88, 0.875, 0.799, and 0.752 respectively. Consequently, the accuracy of cover estimation was highest for the camera and lowest for the ceptometer. The sighting-frame gave better cover estimates than the metre- stick. Precision was highest for the camera and ceptometer. This was attributed to their high objectivity contrary to the sighting frame and metrestick which are more exposed to subjective bias. Analysis of variance showed that all the techniques were independent and there was no interaction between any technique and the crop. The sighting frame was however recommended for routine field application for row crops due to its cost effectiveness, local availability, ease of use and reI iable accuracy. Two approaches in the use of the Sighting frame were also compared; the conventional 0-0.5-1 and the new 0-0.25-0.5-0.75-1 systems. Both approaches had r-values in the range of 0.95 - 0.99. Consequently any of the two systems could be adopted for routine cover evaluation. A preliminary rainfall energy interception model for the short rains 1992/93 at the research site was also suggested. Of importance were the effects of extreme rainfall events realised towards late 1992 and early 1993. It was recommended that for all cover measurement techniques, evaluation procedures standardization and design improvements be made to increase the precision and accuracy of cover estimation and to facilitate ease of application in the field.