The water table in southwestern Niger has been rising continuously for the past decades (4 m rise from 1963 to 2007), despite a ∼23% deficit in monsoonal rainfall from 1970 to 1998. This paradoxical phenomenon has been linked with a change in land use from natural savannah to millet crops that have expanded in area sixfold since 1950 and have caused soil crusting on slopes that has, in turn, enhanced Hortonian runoff. Runoff concentrates in closed ponds and then recharges the aquifer; therefore, higher runoff increases aquifer recharge. At the local scale (2 km2), a physically based, distributed hydrological model showed that land clearing increased runoff threefold, whereas the rainfall deficit decreased runoff by a factor of 2. At a larger scale (500 km2, 1950–1992 period), historical aerial photographs showed a 2.5‐fold increase in the density of gullies, in response to an 80% decrease in perennial vegetation. At the scale of the entire study area (5000 km2), analytical modeling of groundwater radioisotope data (3H and 14C) showed that the recharge rate prior to land clearing (1950s) was about 2 mm a−1; postclearing recharge, estimated from groundwater level fluctuations and constrained by subsurface geophysical surveys, was estimated to be 25 ± 7 mm a−1. This order of magnitude increase in groundwater fluxes has also impacted groundwater quality near ponds, as shown by a rising trend in groundwater nitrate concentrations of natural origin (75% of δ15N values in the range +4 to +8‰). In this well‐documented region of semiarid Africa, the indirect impacts of land use change on water quantity and quality are much greater than the direct influence of climate variability.

No abstract

Drought-induced tree mortality is occurring across all forested continents and is expected to increase worldwide during the coming century. Regional-scale forest die-off influences terrestrial albedo, carbon and water budgets, and landsurface energy partitioning. Although increased temperatures during drought are widely identified as a critical contributor to exacerbated tree mortality associated with "global-change-type drought", corresponding changes in vapor pressure deficit (D) have rarely been considered explicitly and have not been disaggregated from that of temperature per se. Here, we apply a detailed mechanistic soil-plant-atmosphere model to examine the impacts of drought, increased air temperature (+2°C or +5°C), and increased vapor pressure deficit (D; +1 kPa or +2.5 kPa), singly and in combination, on net primary productivity (NPP) and transpiration and forest responses, especially soil moisture content, leaf water potential, and stomatal conductance. We show that increased D exerts a larger detrimental effect on transpiration and NPP, than increased temperature alone, with or without the imposition of a 3-month drought. Combined with drought, the effect of increased D on NPP was substantially larger than that of drought plus increased temperature. Thus, the number of days when NPP was zero across the 2-year simulation was 13 or 14 days in the control and increased temperature scenarios, but increased to approximately 200 days when D was increased. Drought alone increased the number of days of zero NPP to 88, but drought plus increased temperature did not increase the number of days. In contrast, drought and increased D resulted in the number of days when NPP = 0 increasing to 235 (+1 kPa) or 304 days (+2.5 kPa). We conclude that correct identification of the causes of global change-type mortality events requires explicit consideration of the influence of D as well as its interaction with drought and temperature.

No abstract