Planning of Optimized Irrigation Decision in Weather to Extended Range
using Weather Forecast with a Coupled Framework of Optimization and
Ecohydrological Model
Abstract
Optimization in irrigation scheduling using weather forecast has been
proven to achieve better productivity along with reduced irrigation
water requirements. We developed a farm-scale hydrological model coupled
with a chance-constraint optimization to take short to medium range
weather forecast and prescribe the optimal irrigation amount determined
by developing the conditional probability density functions of the
rainfall and subsequently the soil moisture for the days in forecast
range. The stress-avoidance was ensured by maintaining the probability
of crops undergoing water stress is less than a prescribed threshold
(reliability factor, α). The framework was implemented for irrigation
decision simulation at extended range by downscaling the forecast with
Nonhomogeneous Hidden Markov Model (NHMM) as an input and produce
irrigation decision in extended range (15 to 30 days). The optimization
framework ensured minimal water use without significant crop water
stress. The method was tested at two site locations in Nashik district
in the state of Maharashtra, both being involved in grape cultivation
(referred herein as Site 1 and Site 2). In short-to-medium range weather
scale, the model was implemented with varied α (0.5 to 0.95) and
interval between two subsequent irrigation application (1, 3 and 7 days)
and significant amount of water savings with respect to the farmer’s
applied irrigation could be achieved. The simulation-optimization
framework was only tested with α=0.95 and once in 7 days irrigation
application for extended range, and yet no significant detrimental
effect on yield was observed whereas in kharif season significant
potential of water savings was observed both in Site 1 and 2. While the
framework in short to medium range is useful for optimal real time
irrigation decision making, in the extended range, it can be implemented
in planning of irrigation for the upcoming month to avoid the
inconvenience of instant arrangement of water, especially in case of
drought-hit regions. Considering that irrigation accounts for over 80%
of the total water use worldwide, the value of such an approach as a
decision-support tool for irrigation optimization is self-evident.