Changes in soil hydraulic conductivity and preferential flow pathways after assisted forest restoration on degraded land in the Khasi Hills (Meghalaya, NE India)

As in other parts of the Indian subcontinent, the forests of Meghalaya (NE India) provide an array of environmental services but the prevalence of traditional slash-and-burn agriculture (locally called jhum) plus mining has led to severely degraded hillsides and a critical loss of soil water storage opportunity and groundwater recharge. As a result, despite receiving some of the highest rainfall totals in the world (MAP up to 11 m, 75% received between May and September), the Meghalaya plateau faces severe water scarcity during the five-month long dry season. In response to such problems, initiatives have been taken towards restoring hydrological functioning through reforestation and assisted natural regeneration (ANR) programmes. As a first step towards assessing the possible improvement of soil physical characteristics and associated hydrological functioning after several years of ANR we measured topsoil (0–10 cm) saturated hydraulic conductivity (K_sat) using double-ring infiltrometry at 12 sites in the Khasi Hills that represented three contrasting vegetation covers: (i) sacred forest (n = 6, natural baseline), (ii) 2–10-year-old ANR (n = 3), and (iii) degraded Imperata grassland (n = 3, degraded reference). At each site, nine K_sat-measurements were taken along the hillslope gradient. In addition, at three sites, blue dye infiltration experiments (n = 2 per site) were carried out to examine the dominant percolation pathways associated with each land-cover type. The median K_sat value for the sacred forest sites was highest (373 mm h^-1), reflecting the abundance of biologically mediated macropores arising from the decomposing activity of soil microflora and fauna at these relatively undisturbed sites. The corresponding value for the ANR sites (160 mm h^-1) was much higher than the median K_sat for the degraded grasslands (71 mm h^-1) but still considerably below the forest reference. Limited observations of topsoil bulk density and carbon content (n = 5 samples in each of three plots) showed increasing bulk density and decreasing carbon content from forest via ANR to grassland, thereby reflecting the observed trend in K_sat. The blue dye experiments suggested infiltration in the sacred forest was dominated by flow along roots and other preferential flow pathways whereas the degraded grassland was mostly characterized by matrix flow. The ANR site showed intermediate behaviour with macropore flow exhibiting high matrix interaction. Comparison of observed median topsoil K_sat in top-layer with prevailing (maximum) hourly rainfall intensities for Cherrapunji suggested infiltration-excess overland flow (IOF) must be considered a rare phenomenon in the sacred forest. Conversely, the K_sat-values for the ANR and degraded grassland sites indicated the occurrence of IOF at high-intensity rainfall events. Despite the observed improvement in surface K_sat it cannot be excluded that the generally shallow nature and high stoniness of the soils pose serious limitations to rebuilding soil water storage capacity through ANR/reforestation. Furthermore, frequent occurrence of saturation-excess OF at the height of the monsoon and associated surface erosion cannot be excluded.