Precipitation regimes are predicted to become more variable with more extreme rainfall events punctuated by longer intervening dry periods. Water-limited ecosystems are likely to be highly responsive to altered precipitation regimes. The bucket model predicts that increased precipitation variability will reduce soil moisture stress and increase primary productivity and soil respiration in aridland ecosystems. To test this hypothesis, we experimentally altered the size and frequency of precipitation events during the summer monsoon (July through September) in 2007 and 2008 in a northern Chihuahuan Desert grassland in central New Mexico, USA. Treatments included (1) ambient rain, (2) ambient rain plus one 20 mm rain event each month, and (3) ambient rain plus four 5 mm rain events each month. Throughout two monsoon seasons, we measured soil temperature, soil moisture content (y), soil respiration (R s ), along with leaf-level photosynthesis (A net ), predawn leaf water potential (C pd ), and seasonal aboveground net primary productivity (ANPP) of the dominant C 4 grass, Bouteloua eriopoda. Treatment plots receiving a single large rainfall event each month maintained significantly higher seasonal soil y which corresponded with a significant increase in R s and ANPP of B. eriopoda when compared with plots receiving multiple small events. Because the strength of these patterns differed between years, we propose a modification of the bucket model in which both the mean and variance of soil water change as a consequence of interannual variability from 1 year to the next. Our results demonstrate that aridland ecosystems are highly sensitive to increased precipitation variability, and that more extreme precipitation events will likely have a positive impact on some aridland ecosystem processes important for the carbon cycle.

Primary production, a key regulator of the global carbon cycle, is highly responsive to variations in climate. Yet, a detailed, continental-scale risk assessment of climate-related impacts on primary production is lacking. We combined 16 years of MODIS NDVI data, a remotely sensed proxy for primary production, with observations from 1218 climate stations to derive values of ecosystem sensitivity to precipitation and aridity. For the first time, we produced an empirically-derived map of ecosystem sensitivity to climate across the conterminous United States. Over this 16-year period, annual primary production values were most sensitive to precipitation and aridity in dryland and grassland ecosystems. Century-long trends measured at the climate stations showed intensifying aridity and climatic variability in many of these sensitive regions. Dryland ecosystems in the western US may be particularly vulnerable to reductions in primary production and consequent degradation of ecosystem services as climate change and variability increase in the future.

Climate models suggest that extreme rainfall events will become more common with increased atmospheric warming. Consequently, changes in the size and frequency of rainfall will influence biophysical drivers that regulate the strength and timing of soil CO2 efflux – a major source of terrestrial carbon flux. We used a rainfall manipulation experiment during the summer monsoon season (July–September) to vary both the size and frequency of precipitation in an arid grassland 2 years before and 2 years after a lightning‐caused wildfire. Soil CO2 efflux rates were always higher under increased rainfall event size than under increased rainfall event frequency, or ambient precipitation. Although fire reduced soil CO2 efflux rates by nearly 70%, the overall responses to rainfall variability were consistent before and after the fire. The overall sensitivity of soil CO2 efflux to temperature (Q10) converged to 1.4, but this value differed somewhat among treatments especially before the fire. Changes in rainfall patterns resulted in differences in the periodicity of soil CO2 efflux with strong signals at 1, 8, and 30 days. Increased rainfall event size enhanced the synchrony between photosynthetically active radiation and soil CO2 efflux over the growing season before and after fire, suggesting a change in the temporal availability of substrate pools that regulate the temporal dynamics and magnitude of soil CO2 efflux. We conclude that arid grasslands are capable of rapidly increasing and maintaining high soil CO2 efflux rates in response to increased rainfall event size more than increased rainfall event frequency both before and after a fire. Therefore, the amount and pattern of multiple rain pulses over the growing season are crucial for understanding CO2 dynamics in burned and unburned water‐limited ecosystems.

Measuring environmental variables at appropriate temporal and spatial scales remains an important challenge in ecological research. New developments in wireless sensors and sensor networks will free ecologists from a wired world and revolutionize our ability to study ecological systems at relevant scales. In addition, sensor networks can analyze and manipulate the data they collect, thereby moving data processing from the end user to the sensor network itself. Such embedded processing will allow sensor networks to perform data analysis procedures, identify outlier data, alter sampling regimes, and ultimately control experimental infrastructure. We illustrate this capability using a wireless sensor network, the Sensor Web, in a study of microclimate variation under shrubs in the Chihuahuan Desert. Using Sensor Web data, we propose simple analytical protocols for assessing data quality “on‐the‐fly” that can be programmed into sensor networks. The ecological community can influence the evolution of environmental sensor networks by working across disciplines to infuse new ideas into sensor network development.

BioOne Complete (complete.BioOne.org) is a full-text database of 200 subscribed and open-access titles in the biological, ecological, and environmental sciences published by nonprofit societies, associations, museums, institutions, and presses.