Drought impacts on water resources system are quantified by a variety of methods since there are various definitions on drought. Therefore, there is no consensus on evaluation of duration and severity of the drought (Wilhite and Glantz, 1985). In general, droughts are defined by the variable to be assessed (Wilhite and Glants, 1985; American Meteorological Society, 1997). Among various drought indexes, meteorological drought index has been widely used due to its straightforward definition and easy data accessibility. SPI (Standardized Precipitation Index, McKee et al., 1993) is a well-known meteorological drought index characterizing lack of precipitation on a range of timescales (Lorenzo-Lacruz et al., 2010; Zhai et. al., 2010).
On the other hand, SDI (Streamflow Drought Index) has been used as a hydrologic drought index, which is based on the same calculation methodology with the SPI. SDI is defined as a shortage of streamflow comparing to its normal condition (Tallaksen, 1997; Tallaksen and Van Lanen, 2004; Sung and Chung, 2014). To define drought conditions, i.e., to judge whether there is lack of streamflow, the proper value for threshold is usually investigated. Here, the threshold level is generally determined using flow-duration curve of seasonal, monthly, and daily scale (Sung and Seo, 2018). Thus, drought condition is defined as a period when the streamflow is less than the threshold value. For drought management, SDI analysis is preferred by the regions where streamflow withdrawal dominates the total water supply, because the shortage of streamflow directly leads to lack of water supply. Hence, drought management plan for those regions where streamflow is dominant source for water supply should be made considering hydrologic drought information.
Water shortage in not only affected by natural variability but also anthropogenic impacts such as water diversions and withdrawals. When it comes to agricultural water use, since persistent meteorological drought also leads to decrease in soil moisture, dependence on streamflow withdrawal increases in order to compensate lack of precipitation. Then, it eventually causes further decrease in streamflow. To cut off this risk of water shortage, a soft-path solution, which includes smart technologies and policies that focus on optimal operation and efficiency of water supply system, has been widely used (Brooks and Holtz, 2009). The soft-path solution (Olmstead and Stavins, 2009) aims to improve the overall efficiency of water supply rather than adopting new sources (Sankarasubramanian et al., 2017; Das Bhowmik et al., 2020). For example, it delivers diverse water services to meet water demand at a community level (Gleick et al., 2002).
The soft-path solution has various uncertainty sources which can be reduced by the data collection and regarding analysis. The prediction of water availability is one of the critical tasks in water management because proper water use strategy has been required for sustainable water use especially under intense drought. The water use can be a key variable in the prediction of water availability. In the case of agricultural use, it has been well-known that the water use data are strongly correlated with meteorological variables. Under the below-normal condition of water availability in such as rivers, lakes and reservoirs, it can cause severe water shortage with high water demand. Then it eventually threatens water supply and associated food production (Falkenmark et al 1997, Döll et al 2009, Wisser et al 2010).
Therefore, this study aims to identify the relationships between meteorological variables and irrigation water use rates based on an assumption that water withdrawal for agricultural use would be affected by natural weather conditions. For this purpose, this study explores the relationship between drought conditions and the amount of irrigation water withdrawal from the river. Various drought indicators such as meteorological drought index, potential evapotranspiration (PET) and Normalized difference vegetation index (NDVI) are investigated in this study.