The episodic nature of water availability in arid and semiarid ecosystems has significant consequences on belowground carbon and nutrient cycling. Pulsed water events directly control belowground processes through soil wet-dry cycles. Rapid soil microbial response to incident moisture availability often results in almost instantaneous C and N mineralization, followed by shifts in C/N of microbially available substrate, and an offset in the balance between nutrient immobilization and mineralization. Nitrogen inputs from biological soil crusts are also highly sensitive to pulsed rain events, and nitrogen losses, particularly gaseous losses due to denitrification and nitrate leaching, are tightly linked to pulses of water availability. The magnitude of the effect of water pulses on carbon and nutrient pools, however, depends on the distribution of resource availability and soil organisms, both of which are strongly affected by the spatial and temporal heterogeneity of vegetation cover, topographic position and soil texture. The 'inverse texture hypothesis' for net primary production in water-limited ecosystems suggests that coarse-textured soils have higher NPP than fine-textured soils in very arid zones due to reduced evaporative losses, while NPP is greater in fine-textured soils in higher rainfall ecosystems due to increased water-holding capacity. With respect to belowground processes, fine-textured soils tend to have higher water-holding capacity and labile C and N pools than coarse-textured soils, and often show a much greater flush of N mineralization. The result of the interaction of texture and pulsed rainfall events suggests a corollary hypothesis for nutrient turnover in arid and semiarid ecosystems with a linear increase of N mineralization in coarse-textured soils, but a saturating response for fine-textured soils due to the importance of soil C and N pools. Seasonal distribution of water pulses can lead to the accumulation of mineral N in the dry season, decoupling resource supply and microbial and plant demand, and resulting in increased losses via other pathways and reduction in overall soil nutrient pools. The asynchrony of resource availability, particularly nitrogen versus water due to pulsed water events, may be central to understanding the consequences for ecosystem nutrient retention and long-term effects on carbon and nutrient pools. Finally, global change effects due to changes in the nature and size of pulsed water events and increased asynchrony of water availability and growing season will likely have impacts on biogeochemical cycling in water-limited ecosystems.

Field manipulative experiments represent a straightforward way to explore temporal relationships between annual precipitation and productivity. Water exclusion usually involves the use of rainout shelters, which are in general formed by a complete roof that intercepts 100% of the rainfall and require complicated mechanisms to move the shelter into place. The rainout-shelter design described here is a fixed-location shelter with a roof consisting of bands of transparent acrylic that blocks different amounts of rainfall while minimally affecting other environmental variables. We constructed thirty 3.76-m shelters in an arid steppe near Río Mayo, Argentina (at 45°41'S, 70°16'W), to impose 30%, 55%, and 80% of rainfall interception. We tested the effectiveness of the design by collecting all the intercepted water in storage tanks and we evaluated changes in soil water content with the time domain reflectometry technique. We also measured soil water content in regular grids to assess the magnitude of the edge effect. We analysed the microclimate impact of the shelters by measuring photosynthetically active radiation and air and soil temperature inside and outside shelters. We did not detect significant differences between the observed and the expected rainfall interception for the 30% and 55% interception treatments but the 80% shelters intercepted 71% of incoming rainfall, which was significantly (P<0.05) lower than the expected value. Soil water content was significantly (P<0.05) higher in the control plots than in the plots with rainout shelter at all dates, except in January (summer). Radiation was not affected by the 30% interception treatment, but the roof with the largest number of acrylics bands (80% interception treatment) reduced incident radiation throughout the day by 10%. Air and soil temperatures were lower under than outside the shelters during the period of highest radiation but the opposite occurred with low radiation but with smaller differences. The two characteristics of the shelter, fixed design and low cost, allow for proper replication in space, which is required in ecological field experiments.