The role of macropore flow from plot to catchment scale : a study in a semi-arid area

Desertification and landscape degradation is a worldwide problem, which is expected to grow in time due to unsustainable land use and climate change. In view of these problems, knowledge of the interaction between vegetation, soil moisture and surface runoff, with subsequent erosion risk is essential. This requires mapping of the spatial and temporal variability of infiltration and runoff production. The influence of preferential flow thereupon is nowadays widely recognized. Therefore the role of preferential flow from plot scale to catchment scale on the hydrology is investigated, using field measurements and model applications. The state of the art on preferential flow is described including process descriptions, measurement methods and modeling concepts. Next , the research catchment “Parapu#8221; (approximately 1 km2) in the Extremadura (Spain) is presented and an overview of the used measurement methods is given. The measurement results show that preferential flow occurs throughout the catchment and at different scales. An elaborate analysis of the soil moisture content development, piezometer water-levels and rainfall – discharge relationships, leads to the conclusion that a large network of connected macropores exists. Rapid vertical as well as lateral flow through this network occurs, regardless of the matrix soil moisture conditions. This subsurface flow can contribute from 13% (under wet catchment conditions) up to 80% (under dry catchment conditions) of total discharge. The spatial variability in preferential flow at the plot scale is deduced from dye-tracer infiltration experiments at 18 locations throughout the research catchment. Four characteristics of the stained infiltration patterns are described: the uniform infiltration depth, the maximum depth, the total stained area and the fraction of preferential infiltration. Per plot a large set of additional measurements were performed: porosity, bulk density, saturated conductivity, texture, vegetation, slope, stoniness, and geographical location. Using multiple regression, four of these site variables were found to explain most (50 - 66%) of the spatial variability in preferential flow: texture, vegetation, slope and geographical location. These site variables were used to generate a catchment map with preferential flow characteristics. For local scale infiltration and runoff simulation, we used the ecohydrological model SWAP. Three of the infiltration plots were used to determine SWAP macropore parameters using inverse modeling. The infiltration patters can generally not be reproduced in the simulations without the macropore concept. Also for water balance modeling under natural circumstances, the macropore model leads to slightly better results for infiltration and surface runoff behavior, than the model with an “effective” measured average saturated conductivity. For catchment scale modeling, the 3D macropore model Hillflow is used, which fits the conceptual behavior of the catchment. The results of this model are as good as the results of modeling without a macropore concept, but scenario runs with both models show large differences in rainfall- discharge relationships. This is a strong argument for the use of the model which better fits the real hydrological processes in the research area