Determining the influence of agricultural land use on climate variables for the Canadian Prairies.

Kaharabata, S.K., Banerjee, S.M., Kieser, M., Desjardins, R.L., and Worth, D.E. (2014). "Determining the influence of agricultural land use on climate variables for the Canadian Prairies.", International Journal of Climatology, 34(15), pp. 3849-3862. doi : 10.1002/joc.3946  Access to full text


Since the 1970s the Canadian Prairies have experienced a steady replacement of summer fallow with continuous cropping practices. It has been suggested that such regional scale changes in agricultural land use result in the modification of the regional climate and its variability. Our study explores this influence using linear discriminant analysis (LDA) to determine whether agricultural land use variables can discriminate the classes of exceedance (positive/negative beyond a given small scale variability range) found in the maximum and minimum daily temperature, diurnal temperature range (DTR), daily precipitation and daily incoming solar radiation. These classes of exceedance describe whether the climate was warmer/cooler, drier/wetter and sunnier/cloudier relative to the small scale variability range. The land use data consisted of summer fallow, annuals and perennials for the 127 ecodistricts in the Prairies ecozone. The climate data were neighbouring station data interpolated to the ecodistrict's centroid. Both data sets covered the period 1951–2006 and the analysis covered the entire data period with focus on 1971–2006 that coincided with the decline in summer fallow practice. It considered the primary growing season months of June, July and August and used the winter months of January and February as a contrast. Our approach showed that where large scale changes in agricultural land cover occurs we can establish a statistical relationship between such changes and the modification of the climate's variability. June coincided with the period of maximum positive growth rate and hence land cover change. The greatest discriminating power was therefore found to occur in June for the exceedance classes of maximum and minimum temperature and solar radiation over the 1971–2006 period with success rates ranging from 85 to 93%. The greatest success in discriminating the DTR classes was at 82% in July.

Date modified: