Building the vegetation drought response index for Canada (VegDRI-Canada) to monitor agricultural drought: first results

Abstract

Drought is a natural climatic phenomenon that occurs throughout the world and impacts many sectors of society. To help decision-makers reduce the impacts of drought, it is important to improve monitoring tools that provide relevant and timely information in support of drought mitigation decisions. Given that drought is a complex natural hazard that manifests in different forms, monitoring can be improved by integrating various types of information (e.g., remote sensing and climate) that is timely and region specific to identify where and when droughts are occurring. The Vegetation Drought Response Index for Canada (VegDRI-Canada) is a recently developed drought monitoring tool for Canada. VegDRI-Canada extends the initial VegDRI concept developed for the conterminous United States to a broader transnational coverage across North America. VegDRI-Canada models are similar to those developed for the United States, integrating satellite observations of vegetation status, climate data, and biophysical information on land use and land cover, soil characteristics, and other environmental factors. Collectively, these different types of data are integrated into the hybrid VegDRI-Canada to isolate the effects of drought on vegetation. Twenty-three weekly VegDRI-Canada models were built for the growing season (April–September) through the weekly analysis of these data using a regression tree-based data mining approach. A 15-year time series of VegDRI-Canada results (s to 2014) was produced using these models and the output was validated by randomly selecting 20% of the historical data, as well as holdout year (15% unseen data) across the growing season that the Pearson’s correlation ranged from 0.6 to 0.77. A case study was also conducted to evaluate the VegDRI-Canada results over the prairie region of Canada for two drought years and one non-drought year for three weekly periods of the growing season (i.e., early-, mid-, and late season). The comparison of the VegDRI-Canada map with the Canadian Drought Monitor (CDM), an independent drought indicator, showed that the VegDRI-Canada maps depicted key spatial drought severity patterns during the two targeted drought years consistent with the CDM. In addition, VegDRI-Canada was compared with canola yields in the Prairie Provinces at the regional scale for a period from 2000 to 2014 to evaluate the indices’ applicability for monitoring drought impacts on crop production. The result showed that VegDRI-Canada values had a relatively higher correlation (i.e., r > 0.5) with canola yield for nonirrigated croplands in the Canadian Prairies region in areas where drought is typically a limiting factor on crop growth, but showed a negative relationship in the southeastern Prairie region, where water availability is less of a limiting factor and in some cases a hindrance to crop growth when waterlogging occurs. These initial results demonstrate VegDRI-Canada’s utility for monitoring drought-related vegetation conditions, particularly in drought prone areas. In general, the results indicated that the VegDRI-Canada models showed sensitivity to known agricultural drought events in Canada over the 15-year period mainly for nonirrigated areas.

Keywords:

• Drought monitoring
• crop yield
• vegetation condition
• Canadian prairies

Highlights

A new Vegetation Drought Response Index for Canada (VegDRI-Canada) is developed to monitor for Canada.

A case study was conducted to evaluate the VegDRI-Canada results over the prairie region of Canada for two drought years and one non-drought year for three weekly periods of the growing season.

Results show that the VegDRI-Canada models showed sensitivity to known agricultural drought events in Canada over the 15-year period mainly for nonirrigated areas.

Acknowledgments

This study was supported by Agriculture and Agri-Food Canada (AAFC) and the National Drought Mitigation Center (NDMC), University of Nebraska–Lincoln. J. Brown was supported by the Land Remote Sensing Program in the USGS Climate and Land Use Change mission area. The authors thank Karin Callahan and Chis Poulsen at the NDMC for extracting the satellite and biophysical data for the model development. We would also like to acknowledge the expertise of Canadian scientists Bahram Daneshfar and Yinsuo Zhang, who contributed to this study. We also thank Deborah Wood of the NDMC for her editorial comments. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was supported by Agriculture and Agri-Food Canada (AAFC) and the National Drought Mitigation Center (NDMC), University of Nebraska–Lincoln.