Relationships between landscape pattern, wetland characteristics, and water quality in agricultural catchments
Water quality in streams is dependent on landscape metrics at catchment and wetland scales. A study was undertaken to evaluate the correlation between landscape metrics, namely patch density and area, shape, heterogeneity, aggregation, connectivity, land-use ratio, and water quality variables (salinity, nutrients, sediments, alkalinity, other potential pollutants and pH) in the agricultural areas of a semiarid Mediterranean region dominated by irrigated farmlands (NE Spain). The study also aims to develop wetland construction criteria in agricultural catchments. The percentage of arable land and landscape homogeneity (low value of Simpson index) are significantly correlated with salinity (r2 = 0.72) and NO3–N variables (r2 = 0.49) at catchment scale. The number of stock farms was correlated (Spearman's corr. = 0.60; p < 0.01) with TP concentration in stream water. The relative abundance of wetlands and the aggregation of its patches influence salinity variables at wetland scale (r2 = 0.59 for Na+ and K+ concentrations). The number and aggregation of wetland patches are closely correlated to the landscape complexity of catchments, measured as patch density (r2 = 0.69), patch size (r2 = 0.53), and landscape heterogeneity (r2 = 0.62). These results suggest that more effective results in water quality improvement would be achieved if we acted at both catchment and wetland scales, especially reducing landscape homogeneity and creating numerous wetlands scattered throughout the catchment. A set of guidelines for planners and decision makers is provided for future agricultural developments or to improve existing ones.