ABSTRACT
Secondary forests and coffee agroecosystems are considered good alternatives for conservation of a high capacity for water filtration in the soil where tropical montane cloud forests (TMCF) once grew; however, it is not clear which characteristics of the vegetation modulate the field saturated hydraulic conductivity of the soil (Kfs) and whether these characteristics persist in such derived systems. Here, we explore how changes in vegetation between secondary forests and coffee agroecosystems have consequences for the average value and spatial variation of litter thickness and root biomass, and whether these differences can explain the Kfs and its spatial distribution. We found that the thickest litter, greatest total biomass and thickest roots are in the secondary forest of the north of the study area. The litter is spatially structured in patches of ca. 12m at plot scale in the secondary forest and coffee agroecosystem of the southern area. Like the Kfs, the thickness of the litter and biomass of the thick (>2mm), medium (1-2mm) and fine (<1mm) roots are spatially distributed on a north to south gradient at landscape scale. Our linear model indicates that geographic area (north or south), land use and litter thickness explain the Kfs and its spatial distribution along this gradient. Even on inclusion of the antecedent soil moisture and percentage of clays (found to explain Kfs in a previous study), it was not possible to eliminate from the model geographic area and land use, due to their high explanatory power. However, antecedent soil moisture became redundant on inclusion of the litter layer, which had a greater explanatory power. Our modeling suggests that undiscovered differences prevail between the geographic areas and secondary forest and coffee agroecosystems (possibly related to the edaphogenesis and management practices) that determine the Kfs.
