Investigation of Climate Change Adaptation Impacts on Optimization of Water Allocation Using a Coupled SWAT-bi Level Programming Model

Abstract

One way to deal with the future effects of climatic changes on the water resources and to cope with water shortages in basins is to have a clear understanding of the future climate change trends. To this end, this study proposes an integrated hydrological-water transfer and supply (HWTS) framework including a coupled SWAT-Bi level programming model to investigate future optimal water supply between different sectors with regard to transaction right. Indeed, Soil & Water Assessment Tool (SWAT) is applied to project the rate of streamflow under Representative Concentration Pathway scenarios of RCP2.6&RCP4.5 and future periods (2020–2040 & 204(Abbas et al. 2015)2060). In addition, a case study of the Hamoun wetland in southeastern of Iran is considered for calibration and validation of real historical data (2000–2016) and then simulation of future streamflow patterns (2020–2060). Next, simulated streamflow data extracted by SWAT is entered as the input of market based bi-level optimization model so that upper-level manager seeks optimization of the available water level in the reservoirs while the lower-level decision maker tries to minimize the economic loss due to water shortage between different sectors regarding transaction right. However, after solving the model with the Improved Particle Swarm Optimization (IBPSO) technique, the final results show that although not much economic profit will be made, but considering specific management strategies such as demand reduction schemes to conserve more water, the imbalance between supply and demand can be significantly improved.

Supplementary materials

Stackelberg (Leader-Follower) framework

Stackelberg game (leader-follower) framework, involves a decision-making process with two decision-makers in view of a bi-level model with hierarchical structure (Guo et al. 2012; Yao et al. 2019), in which the leader of the upper level has the freedom to choose the best decision according to the reaction of the follower of the lower level, so he optimizes the objective function related to his level minF(X, Y) by considering the relevant level constraints U(X, Y) ≤ 0, and on the other hand, the follower according to the decision of the leader must optimize the objective function of his level minf(x, y) regarding lower-level constraints u(x, y) ≤ 0. The standard formulation for the bi-level model is as follows:

$$ {\displaystyle \begin{array}{c}\underset{x}{\min}\kern0.24em F\left(X,Y\right)\\ {}s.t.\kern0.6em U\left(X,Y\right)\le 0\\ {}\underset{y}{\min}\kern0.36em f\left(x,y\right)\\ {}s.t.\kern0.72em u\left(x,y\right)\le 0\end{array}} $$

(16)

Parameter Selection

Identification of the most sensitive parameters and their precisions for the basin is the first step in the calibration and validation process (Arnold et al. 2012). Parameterizing is a complex process due to the different characteristics of basins such as soil maps, etc. (Pichuka and Maity 2017). In this study, several sensitive parameters (Table 7) with different ranks of sensitivity in the entire basin have been considered in the SWAT-CUP. This is due to the difference in the slope of the region so that the parameters of groundwater and soil are the most sensitive in areas with lower slopes, while runoff parameters in areas with higher slopes are considered as the most sensitive parameters (Schmalz and Fohrer 2009). In fact, in areas with lower slopes, the amount of rain penetration is higher, which leads to an increase in base flow (Fereidoon and Koch 2018). Therefore, the groundwater parameters, GWQMN, ALPHA_BF, RCHRG_DP and SHALLST, followed by the soil parameter SOL_BD, were found to be more sensitive in lower elevations than in higher elevations. It must be mentioned that the CN2 (curve number in SCS method) is the most sensitive parameter for the whole basin.

Table 7 Initial and final parameter ranges for SWAT calibration and parameter significance

Information on the Percentage of each Land Cover in the Hamoun Wetland

Changes in land use classification such as decreasing forest cover, rising urban areas, increasing barren lands, etc., both in the short and long term, have a direct impact on the final output of the projected hydrological patterns in the basin (Worku et al. 2017; Abbas et al. 2015). In fact, in hydrological studies, analyzing the effect of land use change on the water balance of the basin is a priority because such a change affects the type of land cover, leads to changes in quantity, velocity, and intensity of surface flow, and then affects the hydrological process of the basin (Ghaffari et al. 2010; Worku et al. 2017). According to the above description, the land use classification utilized in SWAT for projecting climate change in Hamoun Wetland is as follows (Table 8):

Table 8 Land use classification for Hamoun Wetland

Table 9 Water demand in two states considering different periods (10⁶m³)

Table 10 Parameters for water transfer in HWTS