Objective Measurement of Drought and Agricultural Response to Drought in the Water-Constrained Monsoon Climate of Tamil Nadu, India

Rodgers, Charles; Svendsen, Mark; ,

Objective Measurement of Drought and Agricultural Response to Drought in the Water-Constrained Monsoon Climate of Tamil Nadu, India

by Charles Rodgers, Univ of Wisconsin, Madison, United States,
Mark Svendsen, Univ of Wisconsin, Madison, United States,

Document Type: Proceeding Paper

Part of: Irrigation and Drainage: Saving a Threatened Resource?In Search of Solutions

Abstract:

Existing definitions of drought, particularly those based on annual or seasonal probability of occurrence, can fail to characterize adequately the relationship between transient shortfalls in precipitation and changes in agricultural output where a significant fraction of crop water needs are supplied by irrigation. This paper presents a discussion of the impact of variation in precipitation on aggregate harvested area, yield and output in the State of Tamil Nadu, India over the period 1968-1986. Seasonal demand for water is estimated for 12 important crops, and demand is allocated between direct rainfall and three sources of irrigation: canals, tanks and groundwater. The relative yield impact of each source of irrigation is also estimated. It is found that aggregate yields associated with groundwater irrigation are approximately twice as high as surface irrigation yields. Irrigation from all sources is estimated to supply between 40% and 50% of agricultural water demand on average. It is observed that interannual changes in aggregate crop output are not strongly associated with changes in seasonal rainfall. An index of annual drought status is proposed that is based on the lagged precipitation series. This index reflects the relative importance of each type of stored water utilized as irrigation and the unique lag structure that relates each source of stored water to direct precipitation. In particular, groundwater levels are estimated to possess a precipitation memory of over 20 months. The proposed index is found to be more highly correlated with both harvested area and yield than is current year or current and previous year precipitation.

Return to search