B34A-01 INVITED
Long-Term Trends in River Flow From Semiarid Rangelands: Degradation and Recovery
In the last 100 years or so, desertification, degradation, and woody plant encroachment have altered huge tracts of semiarid rangelands. It is expected that the changes thus brought about significantly affect water balance in these regions; and in fact, at the headwater-catchment and smaller scales, such effects are reasonably well documented. For larger scales, however, there is surprisingly little documentation of hydrological change. In this paper, we evaluate the extent to which streamflow from large watersheds in central Texas has changed concurrent with the dramatic shifts in vegetation cover (conversion from grasslands to woodlands) that have taken place during the last century. Our study focused on the three watersheds that supply the major tributaries of the Concho River—those of the North Concho (3279 km2), the Middle Concho (5398 km2), and the South Concho (1070 km2). Using data from the period of record (1926-2005), we found that annual streamflow for the North Concho decreased by about 70% between 1960 and 2005. Not only did we find no downtrend in precipitation that might explain this reduced flow, we found no corresponding change in annual streamflow for the other two basins. When we analyzed trends in baseflow and stormflow separately, however, we found that in spite of large increases in mesquite and juniper cover, baseflow for all the watersheds has remained essentially consistent since 1960 (and may even have increased). At the same time, stormflows (floods) were fewer and of smaller magnitude. In other words, the same amount of precipitation, on average, generated less stormflow after 1960 than before. We suggest that these lower stormflows result from generally higher soil infiltrability due to the greater density of both woody and herbaceous plants. There is no indication that the decline in streamflow is related to diminished baseflows caused by extraction of subsurface water by woody plants.
B34A-02 INVITED
Ecosystem-groundwater interactions under changing land uses: Linking water, salts, and carbon across central Argentina
Although most ecosystems display a one-way connection with groundwater based on the regulation of deep
water drainage (recharge), this link can become reciprocal when the saturated zone is shallow and plants take up
groundwater (discharge). In what context is the reciprocal link most likely? How is it affected by land use
changes? Has it consequences on salt and carbon cycling? We examine these questions across a precipitation
gradient in the Pampas and Espinal of Argentina focusing on three vegetation change situations (mean annual
rainfall): afforestation of humid (900-1300 mm) and subhumid grassland (700-900 mm/yr of rainfall), annual
cultivation of subhumid grasslands (700-800 mm/yr), and annual cultivation of semiarid forests (500-700 mm).
Humid and subhumid grasslands have shallow (< 5 m deep) groundwater tables that are poorly consumed by
grasses but highly used by planted trees, as evidenced by satellite canopy temperatures, soil moisture and water
table level records, and sapflow measurements. Groundwater contributions enhance carbon uptake in
plantations compared to grasslands as suggested by aboveground biomass measurements and satellite
vegetation indexes from sites with and without access to groundwater. Where rainfall is <1100 mm, grassland
afforestation switches water fluxes to groundwater from positive (net recharge) to negative (net discharge)
causing a salt accumulation process in soils and groundwater that is ultimately limited by the tolerance to salinity
of tree species. Cultivation with corn and soybean can lead to groundwater consumption in the driest belt of
subhumid grassland. Up to five-fold yield increases in lowlands vs. uplands during the driest years indicate a
dramatic impact of groundwater use on carbon uptake and groundwater salinization suggests a recharge-to-
discharge switch. In dry forests groundwater is not accessible (> 15 m deep) and recharge under natural
conditions is null. The establishment of crops, however, triggers the onset of recharge, as evidenced by vadose
zones getting wetter and leached of atmospheric chloride. Cropping may cause water table raises leading to a
two-way coupling of ecosystems and groundwater in the future, as it has been documented for similar settings in
Australia and the Sahel. In the Pampas land use change interacts with groundwater consumption leading to
higher carbon uptake (humid and subhumid grasslands) and salt accumulation (subhumid grasslands). In the
Espinal (semiarid forest) land use change currently involves a one-way effect on groundwater recharge that may
switch to a reciprocal connection if regional water table raises occur. Neglecting the role of groundwater in flat
sedimentary plains can obscure our understanding of carbon and salt cycling and curtail our attempts to sustain
soil and water resources under changing land uses.
http:gea.unsl.edu.ar
B34A-03
Ecohydrological Consequences of Shifts in Grass-to-Woody Plant Dominance in Water Limited Ecosystems
An increase in the representation of woody plants in historic grasslands has been a wide-spread recent phenomena in the drylands of North America. The consequences of this vegetation change for ecosystem services are uncertain and likely related to how soil-plant interactions are influenced by precipitation. Here we compare ecosystem water and carbon dioxide fluxes, determined over four years by eddy covariance, for a grassland, a grassland-shrubland mosaic, and a fully developed woodland to evaluate the relationship between land surface cover and biosphere-atmosphere exchange. Since our system is located in a riparian system, it interacts with vegetation type to accentuate differences in soil water availability and helps us to disentangle with how different carbon cycling components are coupled to the hydrologic cycle. Compared to our grassland, our grassland-shrubland mosaic often uses a similar amount of water throughout a growing season (equivalent values of evapotranspiration), whereas it accumulated a similar amount of carbon from the atmosphere as our fully developed mesquite forest (resulting higher ecosystem water-use efficiency). This pattern depends on seasonal precipitation, where changes in the size-class distribution of rainfall events and season total differentially influence ecosystem respiration and photosynthesis, depending upon woody-plant abundance and woody-plant age. Interestingly the differences in vegetation type also change the relationship between soil carbon cycling and precipitation, where microsites with high resource availability become hot-spots of activity that are important in regulating ecosystem carbon balance. A careful understanding of the complexities of both the plant and soil compartment to vegetation change is important to predicting the consequences of vegetation change on biosphere-atmosphere material and energy exchange.
B34A-04
Ecohydrology underground: how root attributes influence water, carbon, and nutrient cycling
Through the depth and volume of the soil that they explore, plants strongly influence the cycling of water, carbon, and nutrients. This talk combines data from two experimental systems in North and South America to demonstrate such interactions. One is a gradient of sites in the southwestern U.S. where we are examining the ecohydrology associated with woody plant encroachment, including a unique cave system that allows us to measure water uptake by plants meters, even tens of meters, underground. Another is a network of plantations that provides sites with contrasting water, carbon, and nutrient use that we compare to adjacent native vegetation, particularly grasslands. We combine field results from these systems with remote sensing observations and model simulations to examine similar ecohydrological/nutrient interactions in other regions of the world.
B34A-05
Evidence and implications of soil moisture-vegetation feedbacks in semiarid savannas
The effect of climate, soil, and vegetation on water-limited ecosystems, is exerted through the soil moisture dynamics. Unlike abiotic factors (e.g., soil texture and rainfall regime), the control exerted by vegetation composition and structure on soil moisture variability remains poorly understood. A number of field studies in dryland landscapes have found higher soil water contents in vegetated soil patches than in adjacent bare soil, providing a convincing explanation for the observed preferential establishment of grasses and seedlings beneath tree canopies. Thus, because in arid and semiarid ecosystems water is the limiting factor for vegetation, a positive feedback could exist between soil moisture and vegetation dynamics. It is still unclear how the strength of such a feedback would change under different long-term rainfall regimes. To this end, we report field observations from savanna ecosystems located along the rainfall gradient in the Kalahari, where the presence of relatively uniform sandy soils limits the effects of co-varying factors. We found that the strength of the positive vegetation-soil moisture feedback increases with increasing levels of aridity. Because positive feedbacks are often associated with the existence of alternative stable states in the dynamics of a system, we developed a minimalist modeling framework to investigate the effect of this feedback on the ecosystem dynamics. We found that, when the feedback is relatively strong, the system may exhibit two stable states corresponding to vegetated and bare soil conditions. We also show how root and soil moisture distributions affect the stability of and resilience of the vegetated (savanna) states along the Kalahari.
B34A-06
Patterns of soil water infiltration and extraction as a major determinant for vertical root distributions
Plants and their roots are the most important source for organic carbon in the soil. Understanding vertical distributions of soil carbon inputs therefore requires knowing the factors that shape vertical root distributions. Where water availability is limiting plant growth, soil water infiltration patterns are likely to shape vertical root distributions. In arid climates with high temporal variability of precipitation, shallow water infiltration is most common and therefore the most reliable, but most short-lived plant water source and deep infiltration the most rare, most long-lived, but least reliable water source. In contrast, in seasonally dry, semi-arid to sub-humid climates, a seasonal water surplus can accumulate at depth and thus be available through deep roots during the dry season. Based on these differences, it is predicted that rooting depths in arid climates are limited by infiltration depths, while seasonally dry climates should favor deep roots, especially in climates with a pronounced wet-season surplus of water. These predictions were tested using data collected in two global databases. The RPGE database contains 564 vertical root profiles of global ecosystems, while the RSIP database contains shapes and sizes of root systems for 2349 individual plants. Data from both databases supported the prediction of infiltration depths as a major factor limiting rooting depths in dry environments as well as the prediction of very deep roots in seasonally dry climates with reliable wet seasons. The relationship between water infiltration patterns and vertical root distributions was formalized in a Soil Water Infiltration and Extraction Model (SWIEM), which models soil water infiltration through 500 discrete soil layers to a depth of 5 m. Water extraction proceeds from the top down, with extraction depths determined by potential evapotranspiration and the vertical distribution of soil water. The hypothesis that water extraction patterns would be predictive of vertical root distributions was tested for eight root profiles from different biomes. Predicted soil water extraction patterns matched observed vertical root distributions very well, thereby suggesting that water infiltration and extraction patterns are indeed a major determinant for vertical root distributions.
B34A-07
Effects of litter position on mass loss and nitrogen release in the semiarid Patagonian steppe
The patchy distribution of vegetation in arid and semiarid ecosystems results in a mosaic of microsites of soil properties and variable abiotic conditions, including the well-documented "islands of fertility", low nutrient conditions in exposed bare soil and large amounts of standing dead material. I evaluated the relative importance of litter position on mass loss and nutrient release in a variety of realistic litter positions both in vegetated, unvegetated, aerial and buried microsites in a natural semiarid steppe in Patagonia, Argentina. Position demonstrated a highly significant effect on mass loss for all litter types (P<0.0001), but surprisingly, the fastest decomposition occurred in litter that was suspended in aerial positions or buried (k=0.25 and 0.32 year-1, respectively), intermediate values for mass loss of litter in bare soil and in shrub removal patches (k=0.21 and 0.24 year-1, respectively), and markedly slowest decomposition occurring under shrub patches (k = 0.018 year-1). In contrast, nutrient release showed a very different pattern with nutrient immobilization occurring only in shrub and buried microsites while all other positions demonstrated a gradual decrease in nitrogen over time. These results support the idea that abiotic photodegradation may be an important driver affecting carbon losses in litter in positions exposed to solar radiation, while nutrient dynamics appear to be largely biotically mediated and concentrated in photoprotected areas where biotic activity dominates. Global change may differentially affect carbon and nutrient turnover due to the relative importance of abiotic and biotic factors affecting litter decomposition in semiarid ecosystems.
B34A-08
Linkages Between Biotic and Abiotic Belowground Processes in a Mojave Desert Ecosystem: Responses to Experimental Nitrogen and Water Additions
Fine roots play a critical role in nutrient acquisition and water uptake. Yet it is unclear how fine roots in arid environments respond to increased nitrogen deposition and rainfall, two important global change factors in arid lands in the southwestern United States. In addition it is unclear how changes in root activity may impact soil CO2 concentrations, an important parameter affecting carbonate dynamics. We measured fine root length density (RLD) and soil CO2 concentrations for two years in experimentally manipulated plots in a Mojave Desert ecosystem. The study was conducted at the Mojave Global Change Facility located at the Nevada Test Site 60 miles northwest of Las Vegas. The treatments included: 1) three 25 mm water additions during the summer, 2) one nitrogen addition in the fall equivalent to 40 kg per hectare per year, 3) a combined water and nitrogen addition and, 4) untreated controls. Root data were collected using minirhizotron imaging approximately every 90 days underneath shrubs and intershrub areas. Soil CO2 concentrations were collected at the same sampling times and locations at 10, 40 and 90 cm depth using gas wells. The RLD showed clear seasonal patterns with the fastest increase in RLD occurring between February and April. During the winter the increase in RLD was higher underneath shrubs than in intershrub areas but during the summer months increases in RLD were similar under shrubs and in intershrub areas. Water additions slightly increased root mortality during the summer but this increase in mortality was not large enough to cause consistent differences in RLD between control and irrigated plots. Nitrogen addition had no effect on root dynamics in any of the plots. In contrast to RLD, irrigation consistently increased soil CO2 concentrations at all depths during the summer even when roots were not actively growing anymore. We speculate that the increased mortality under irrigation causes increased heterotrophic respiration which may explain increases soil CO2 concentrations during the summer. These higher soil CO2 concentrations may result in increased carbonate dissolution especially in combination with increased soil moisture. Preliminary simulations using the CALGYP model indicate however that this increased dissolution will not result in changes in ecosystem C budgets since carbonates are likely to precipitate again deeper in the soil.