Quantifying Landscape Pattern–Hydrological Process Linkage in Northwest Iran

The enormous heterogeneity and complexity of landscape patterns and their linkage with the hydrological responses have rarely been quantified and cataloged, especially in ungauged regions. This research therefore linked the landscape characteristics to hydrological processes using a newly developed runoff landscape index (RLI) at the watershed scale in Ardabil Province, northwest Iran. First, 11 common landscape metrics were calculated using Fragstats 4.2.1 software. Then, a runoff landscape index (RLI) was developed based on land cover (&lambda:C), soil (&lambda:K), and topography (&lambda:S) factors in 28 watersheds. Correlation and regression analyses were also conducted to determine the relationship between RLI, commonly used landscape metrics, and mean base flow. The spatial variations of all meaningful landscape metrics and RLI were considerable throughout the study watersheds. The mean values of &lambda:C, &lambda:K, and &lambda:S were found to be 2.78 &plusmn: 1.08, 0.50 &plusmn: 0.10, and 1.22 &plusmn: 0.30, respectively. The mean RLI varied from 0.00009 in the Lay Watershed with an area of 19.09 km2 to 0.28 in the Boran Watershed with 10,268.95 km2. The correlation coefficient (r &gt: 0.42: p-value &lt: 0.05) was obtained significantly between RLI and only five landscape metrics, including the largest patch index (LPI), landscape shape index (LSI), landscape division index (DIVISION), splitting index (SPLIT), and Shannon&rsquo:s diversity index (SHDI). In addition, a regression model with R2 of 0.97 and 0.67, respectively, in calibration and validation steps was established between river base flow as the dependent variable and main waterway length, LPI, LSI, SPLIT, modified Simpson&rsquo:s diversity index (MSIDI), and &lambda:S as independent variables. The result confirms the significant interdependence of RLI and landscape characteristics, which can be used to interpret the landscape&rsquo:s dynamic and its effects on hydrological processes.