Stream Discharge Prediction Based upon Land Use Land Cover Change and
Climate Data using Google Earth Engine and Multi Resolution Land
Characteristics
Abstract
Land use and land cover (LULC) benefit hydrology by enhancing water
quality, flood control, and groundwater recharge. Vegetated areas, like
forests and crop land, act as natural filters, reducing pollutants, and
stabilizing soil to prevent erosion. They also absorb and slow runoff,
mitigating floods, and allowing water to infiltrate and recharge
groundwater. LULC influences local climate, affecting precipitation and
evaporation rates crucial to the water cycle, and helps maintain a
stable water supply. Moreover, sustainable LULC protects habitats for
aquatic species, supporting biodiversity and ecosystem health, ensuring
a resilient and balanced hydrological system. There is a need to see the
effect of LULC change on the discharge and nearby stream landcover for
better management. A study was conducted to assess the impact of
historical land cover changes from 2000 and 2023 on the hydrological
response of the Reedy River basin Greenville County, South Carolina
(SC). Land cover changes were analyzed using Landsat satellite images
processed through Google Earth Engine and pre-classified images from the
Multi-Resolution Land Characteristics (MRLC) dataset. The effects of
these changes on discharge patterns were evaluated using discharge data
obtained from the USGS National Water Dashboard. Physically based
parameters for hydrological models were estimated using land cover
change maps, along with precipitation and temperature data, through
multiple regression analysis. The study found that the R
2 value was higher, at 95%, for the MRLC dataset
compared to 68% for the Landsat 7 imagery. Despite the Landsat imagery
showing significant changes (p<0.05), the MRLC dataset exhibited
greater accuracy. Due to the lower accuracy and higher fluctuations
observed with Landsat 7 imagery, the MRLC dataset was selected to
identify nearby landcover for better management of riparian areas,
utilizing the cost distance tool and stream raster in ArcGIS Pro. The
result contributed to identify an area which affect the streams
negatively.