Hydrological drought occurs when the amount of surface and subsurface water fails to meet water demand. Recently, the coping capacity for drought has dramatically increased since the improvement of water management facilities, such as dams, reservoirs, ground-water wells, and irrigation systems. Severe and extreme drought still occurs, however, disabling human efforts to overcome its harmful effects. In order to minimize the adverse impacts of drought, drought early warning systems that produce appropriate and timely drought information and deliver it to decision-makers are required to respond to unfavorable drought conditions. The amount of available water in the regions suffering from insufficient water supply should be assessed and monitored regularly, based on hydrological variables such as runoff, reservoir level, and streamflow. There is a lack of observation data for these variables in the upstream regions of watershed basins and/or in developing counties.

Remote sensing can provide a cost-efficient way of deriving drought information for regions with a lack of observation data for meteorological and hydrological variables. Drought information may be obtained more easily from meteorological data based on water balance rather than hydrological data, which is difficult to estimate. In this study, a method to assess and monitor hydrological drought using remotely sensed precipitation and evapotranspiration estimates was investigated for use in regions with limited observation data. Air temperature data at the 2-m level was estimated using remotely sensed data, then evapotranspiration was estimated from the air temperature, and the correlations between precipitation minus evapotranspiration (P-PET) and streamflow percentiles were examined.

Land Surface Temperature data with a 1km×1km spatial resolution as well as Atmospheric Profile data with a 5km×5km spatial resolution from MODIS (Moderate Resolution Imaging Spectroradiometer) sensor on board the Aqua satellite were used to estimate the monthly maximum and minimum air temperature in South Korea. The estimated 2-m air temperature data, that can be used either corrected, using the CRU (Climate Research Unit) TS3.20 gridded dataset, or uncorrected, showed comparable or smaller MAE or RMSE values (MAE=1.18-1.89°C, RMSE=1.42-2.36°C), compared to existing studies. Evapotranspiration was estimated from the maximum and minimum air temperature using the Hargreaves method and the estimates were compared to MOD16 data from the University of Montana, based on the Penman-Monteith method, showing smaller coefficient of determination values but smaller MAE and RMSE (MAE=6.44-25.66mm/month, RMSE=7.54-29.10mm/month).

Precipitation was obtained from Tropical Rainfall Measuring Mission (TRMM) monthly rainfall data, and the correlations of 1-, 3-, 6-, and 12-month P-PET percentiles with streamflow percentiles were analyzed for the Mankyung-gang watershed basin, Dongjin-gang watershed basin, and Upper Namhan-gang watershed basin in South Korea. The Upper Manhan-gang watershed basin was the most affected region from the 2008∼2009 drought, and the Mankyung-gang and Dongjin-gang watershed basins experienced serious drought during the 2012 spring drought.

The 1-month P-PET percentile during JJA in Dongjin-gang watershed basin (Pearson’s r=0.87, p-value⟨.0001; Kendall’s tau=0.65, p-value=0.0004) and 1-month P-PET percentile during JJA (Pearson’s r=0.89, p-value⟨.0001; Kendall’s tau=0.71, p-value⟨.0001) and SON (Pearson’s r=0.63, p-value⟨.0001; Kendall’s tau=0.47, p-value=0.0006) in the Upper Namhan-gang watershed basin are highly correlated with the streamflow percentile with a 95% confidence level. Since the effect of precipitation in these two basins is especially high, the correlation between the evapotranspiration percentile and streamflow percentile is positive (Pearson’s r=0.51, p-value=0.006; Kendall’s tau=0.38, p-value=0.005).

In the Mankyung-gang watershed basin, the 3-month P-PET percentile during MMA (Pearson’s r=0.67, p-value=0.05; Kendall’s tau=0.39, p-value=0.14) and SON (Pearson’s r=0.69, p-value=0.02; Kendall’s tau=0.49, p-value=0.04), as well as JJA (Pearson’s r=0.7, p-value=0.03; Kendall’s tau=0.51, p-value=0.04) are highly correlated with the streamflow percentile with a 95% confidence level. The correlation of the 1-month P-PET percentile with streamflow percentile is also quite high (Pearson’s r=0.88, p-value⟨.0001; Kendall’s tau=0.69, p-value⟨.0001) during JJA.

Remotely sensed air temperature and evapotranspiration estimates produced relatively small MAE and RMSE values compared to existing studies, and the 3-month P-PET percentile in the Mankyung-gang watershed basin and 1-month P-PET percentile in the Upper Namhan-gang watershed basin showed good correlations with streamflow percentile, not only in summer (JJA) with much precipitation, but also in autumn (SON) which is a more important period for drought assessment and monitoring. These results indicate that remotely sensed P-PET estimates can be used for the assessment and monitoring of hydrological drought. There exists a limitation as far as the coarse spatial resolution of precipitation estimates, which can be resolved by the downscaling of satellite precipitation data. Furthermore, soil moisture stress and vegetation phenology should be considered in the estimation of evapotranspiration if the methodology is applied to other regions.

Remote sensing can be used effectively to estimate air temperature and evapotranspiration for regions with limited observation data. The methodology can be applied to the regions that lack meteorological and hydrological data for cost-efficient assessment and monitoring of hydrological drought. The estimates may be used in the decision-making process to minimize the adverse impacts of hydrological drought. The provision of spatially distributed data with high spatial resolution enables the assessment of drought conditions for each region and the establishment of differentiated measures to cope with drought in developing counties.

The production of remotely sensed estimates of meteorological and hydrological variables, as well as derived drought information that is in need in developing counties is an opportunity for international cooperation. The use of these estimates for upstream basins without observation data suggests a cost-effective and timely method for the assessment and monitoring of hydrological drought.

1. 서론

2. 연구 자료 및 방법

2.1 대상 지역

2.2 자료

2.3 연구 방법론

3. 연구내용

3.1 원격 탐사를 통한 대기 온도 추정

3.2 증발산량의 추정

3.3 유역별 자료 산정

4. 결과 및 토의

4.1 원격 탐사를 통한 대기 온도의 추정

4.2 증발산량의 추정

4.3 P-PET와 유량과의 상관성 분석

4.4 수문학적 가뭄 평가

5. 결론