Search

About 40% of global crop production takes place on irrigated land, which accounts for approximately 20% of the global farmland. The great majority of freshwater consumption by human societies is associated with irrigation, which contributes to a major modification of the global water cycle by enhancing evapotranspiration and reducing surface and groundwater runoff. In many regions of the world irrigation contributes to streamflow and groundwater depletion, soil salinization, cooler microclimate conditions, and altered land-atmosphere interactions. Despite the important role played by irrigation in food security, water cycle, soil productivity, and near-surface atmospheric conditions, its global extent remains poorly quantified. To date global maps of irrigated land are often based on estimates from circa year 2000. Here we apply artificial intelligence methods based on machine learning algorithms to satellite remote sensing and monthly climate data to map the spatial extent of irrigated areas between 2001 and 2015. We provide global annual maps of irrigated land at ≈9km resolution for the 2001-2015 and we make this dataset available online.

In shallow coastal bays where nutrient loading and riverine inputs are low, turbidity, and the consequent light environment are controlled by resuspension of bed sediments due to wind-waves and tidal currents. High sediment resuspension and low light environments can limit benthic primary productivity; however, both currents and waves are affected by the presence of benthic plants such as seagrass. This feedback between the presence of benthic primary producers such as seagrass and the consequent light environment has been predicted to induce bistable dynamics locally. However, these vegetated areas influence a larger area than they footprint, including a barren adjacent downstream area which exhibits reduced shear stresses. Here we explore through modeling how the patchy structure of seagrass meadows on a landscape may affect sediment resuspension and the consequent light environment due to the presence of this sheltered region. Heterogeneous vegetation covers comprising a mosaic of randomly distributed patches were generated to investigate the effect of patch modified hydrodynamics. Actual cover of vegetation on the landscape was used to facilitate comparisons across landscape realizations. Hourly wave and current shear stresses on the landscape along with suspended sediment concentration and light attenuation characteristics were then calculated and spatially averaged to examine how actual cover and mean water depth affect the bulk sediment and light environment. The results indicate that an effective cover, which incorporates the sheltering area, has important controls on the distributions of shear stress, suspended sediment, light environment, and consequent seagrass habitat suitability. Interestingly, an optimal habitat occurs within a depth range where, if actual cover is reduced past some threshold, the bulk light environment would no longer favor seagrass growth.

We investigate the influence of vegetation on river morphological instabilities using an analytical framework. We first discuss the important role of the hydrological (flooding frequency) and biological (vegetation development rate) timescales. As long as the changes in riverbed morphology and vegetation over an interval comprising one flood and one low-flow period are small, we show that it is possible to simplify the description of a vegetated river with non-constant discharge. We propose physically-based and effective (neural) models for the feedback between vegetation and morphodynamics. Physically-based approaches use equations of morphodynamics extended to account for the interplay between flow, sediment and vegetation dynamics. While their foundation is solid, a physically-based description is only feasible for simple vegetation cover (grass to shrubs). For complex vegetated obstacles we present as an alternative effective approaches, explicitly including interactions (local and non-local) between obstacles. We focus on the role of vegetation in the emergence of ridge patterns observed in the presence of an ephemeral flow and correspondingly derive a set of conditions for patterns. (C) 2015 Elsevier Ltd. All rights reserved.

Referring to the analytical definition of water wars, several scholars have coherently argued against the “water leads to war thesis.” There are four arguments that have contributed to successfully dispel the myths of water wars: (a) interstate cooperation prevails over conflict; (b) development of new technologies increases freshwater availability; (c) the intrinsic characteristics of water as a resource do not justify interstate military intervention; (d) virtual water trade provides the opportunity to circumvent local water scarcity. These arguments converge demonstrating that rather than water wars in the future, water peace will prevail. While we agree with these arguments on the low likelihood of future water wars, we find that hydropolitical theories have generally neglected the fact that the conditions for interstate water peace come with high socio‐environmental costs. In particular, the central idea that virtual water trade resolve issues of local water scarcity and therefore reduces tensions and escalation of violence among different countries does not fully take into account the fact that dynamics of transnational water appropriation have serious socio‐environmental impacts on the virtual water exporting countries. To conceptualize this phenomenon we introduce the notion of “hidden socio‐environmental costs of virtual water transfer,” which is understood as a specific form of environmental cost‐shifting. The empirical support to our reasoning comes from the study of transnational large‐scale land acquisitions which represent an expanding phenomenon central in the contemporary global agrarian transformation.

In many tropical forests, where phosphorus (P) is considered a limiting nutrient, atmospheric deposition can contribute significantly to available P. Previous studies have shown that P inputs from atmospheric deposition are enhanced by plant canopies. This effect is explained as the result of increased deposition of P-rich aerosol particles (dry deposition) and fog droplets (fog or “occult” deposition) onto leaf surfaces. Here we studied the importance of fog as a source of P to a P-limited dry tropical forest. Throughout an 80 day period during the dry season when fog is most common, we sampled fog water and bulk precipitation in a clearing and measured leaf wetness and throughfall in an adjacent secondary and mature forest stand. During the study period, total P (PT) concentrations in fog water ranged from 0.15 to 6.40 mg/L, on average fourteenfold greater than PT concentrations in bulk precipitation (0.011 to 0.451 mg/L), and sixfold and sevenfold greater than throughfall PT concentrations in the secondary and mature forest stands, respectively (0.007 to 1.319 mg/L; 0.009 to 0.443 mg/L). Based on leaf area index, the frequency of fog deposition, and amount of water deposited per fog event, we estimate that fog delivers a maximum of 1.01 kg/ha/yr to secondary forest stands and 1.75 kg/ha/yr to mature forest stands, compared to 0.88 kg/ha/yr to secondary forest stands and 1.98 kg/ha/yr to mature forest stands via throughfall (wet + dry deposition) and stemflow. Thus, fog deposition may contribute substantially to available P in tropical dry forests.

© 2016 Elsevier Ltd Agricultural intensification is often seen as a suitable approach to meet the growing demand for agricultural products and improve food security. It typically entails the use of fertilizers, new cultivars, irrigation, and other modern technology. In regions of the world affected by seasonal or chronic water scarcity, yield gap closure is strongly dependent on irrigation (blue water). Global yield gap assessments have often ignored whether the water required to close the yield gap is locally available. Here we perform a gridded global analysis (10 km resolution) of the blue water consumption that is needed annually to close the yield gap worldwide and evaluate the associated pressure on renewable freshwater resources. We find that, to close the yield gap, human appropriation of freshwater resources for irrigation would have to increase at least by 146%. Most study countries would experience at least a doubling in blue water requirement, with 71% of the additional blue water being required by only four crops – maize, rice, soybeans, and wheat. Further, in some countries (e.g., Algeria, Morocco, Syria, Tunisia, and Yemen) the total volume of blue water required for yield gap closure would exceed sustainable levels of freshwater consumption (i.e., 40% of total renewable surface and groundwater resources).

© 2017 The escalating human demand for food, water, energy, fibres and minerals have resulted in increasing commercial pressures on land and water resources, which are partly reflected by the recent increase in transnational land investments. Studies have shown that many of the land-water issues associated with land acquisitions are directly related to the areas of energy and food production. This paper explores the land-water-energy-food nexus in relation to large-scale farmland investments pursued by investors from European countries. The analysis is based on a “resource assessment approach” which evaluates the linkages between land acquisitions for agricultural (including both energy and food production) and forestry purposes, and the availability of land and water in the target countries. To that end, the water appropriated by agricultural and forestry productions is quantitatively assessed and its impact on water resource availability is analysed. The analysis is meant to provide useful information to investors from EU countries and policy makers on aspects of resource acquisition, scarcity, and access to promote responsible land investments in the target countries.

© 2017 by the authors. Demographic growth, changes in diet, and reliance on first-generation biofuels are increasing the human demand for agricultural products, thereby enhancing the human pressure on global freshwater resources. Recent research on the food-water nexus has highlighted how some major agricultural regions of the world lack the water resources required to sustain current growth trends in crop production. To meet the increasing need for agricultural commodities with limited water resources, the water use efficiency of the agricultural sector must be improved. In this regard, recent work indicates that the often overlooked strategy of changing the crop distribution within presently cultivated areas offers promise. Here we investigate the extent to which water in the United States could be saved while improving yields simply by replacing the existing crops with more suitable ones. We propose crop replacement criteria that achieve this goal while preserving crop diversity, economic value, nitrogen fixation, and food protein production. We find that in the United States, these criteria would greatly improve calorie (+46%) and protein (+34%) production and economic value (+208%), with 5% water savings with respect to the present crop distribution. Interestingly, greater water savings could be achieved in water-stressed agricultural regions of the US such as California (56% water savings), and other western states.

Hydraulic redistribution defined as the translocation of soil moisture by plant root systems in response to water potential gradients is a phenomenon widely documented in different climate, vegetation, and soil conditions. Past research has largely focused on hydraulic redistribution in deep tree roots with access to groundwater and/or winter rainfall, while the case of relatively shallow (i.e., ≈1–2 m deep) tree roots has remained poorly investigated. In fact, it is not clear how hydraulic redistribution in shallow root zones is affected by climate, vegetation, and soil properties. In this study, we developed a model to investigate the climate, vegetation, and soil controls on the net direction and magnitude of hydraulic redistribution in shallow tree root systems at the growing season to yearly timescale. We used the model to evaluate the effect of hydraulic redistribution on the water stress of trees and grasses. We found that hydraulic lift increases with decreasing rainfall frequency, depth of the rooting zone, root density in the deep soil and tree leaf area index; at the same time for a given rainfall frequency, hydraulic lift increases with increasing average rainstorm depth and soil hydraulic conductivity. We propose that water drainage into deeper soil layers can lead to the emergence of vertical water potential gradients sufficient to explain the occurrence of hydraulic lift in shallow tree roots without invoking the presence of a shallow water table or winter precipitation. We also found that hydraulic descent reduces the water stress of trees and hydraulic lift reduces the water stress of grass with important implications on tree–grass interactions.

During the past several decades, woody plants have been encroaching into grasslands around the world. This transition in plant dominance is often explained as a state shift in bistable ecosystem dynamics induced by fire-vegetation feedbacks. These feedbacks occur when woody plants are able to displace grasses because of their better access to soil water and light. On the other hand, grasses can displace woody plants because of their ability to increase fire frequency and of the higher susceptibility of woody plants to fire-induced mortality. In this study, we present an ecohydrological framework to investigate the displacement of grasses by woody plants. Considering the effect of lateral root spread and of soil water and light limitations, we found that woody plant encroachment can substantially suppress grass production even without the presence of grazers. Bistable dynamics emerge as a result of the grass-fire feedback for a wide range of rainfall conditions, fire susceptibility, and woody plant growth rates.

It is well documented that deforestation results in an increase in landslide frequency due to the control that forest roots have on slope stability. The loss of forest vegetation leads to a reduction in soil cohesion and a decrease in the shear strength of the soil profile. As a result, the slope becomes more susceptible to landsliding and the return time of landslides decreases. When a landslide removes the soil profile, there may not be adequate time for seedlings to grow and enhance soil stability. In this study, we investigate whether bistable dynamics emerge from the interaction of forest vegetation with the formation and accumulation of colluvial deposits in soil-mantled landscapes. To that end, we develop deterministic and stochastic models of landslide occurrence with a dynamic vegetation component. Results show that bistability exists for the deterministic case for both steep and shallow hollows under event and supply limited conditions. However, for the stochastic case, the randomness of landslide occurrence largely changed the states of the system such that the system only exhibited one stable state, which was the fully vegetated condition. Examining different management practices under stochastic conditions showed that the system eventually recovered; however, management practices influenced the recovery time of the forest. Thus, different management practices could render the land in a state of low vegetation over economically significant time periods.

[1] For some phosphorus (P)-limited ecosystems, vegetation can be sustained by atmospheric P inputs. The ability of the canopy to trap airborne particles influences atmospheric P deposition. This dependence suggests a positive feedback, which could impact forest regeneration following deforestation. We examine how the amount of P deposited atmospherically varies as a function of forest canopy characteristics. We quantify the amount of P in bulk deposition (i.e., rainfall and dry deposition) and throughfall from a mature forest and 6 year old successional vegetation stand. To rule out the possibility that P enrichment in throughfall is due to canopy leaching, we construct an artificial forest made of P-free plastic materials. We then compare throughfall samples collected beneath the artificial forest with those collected beneath the successional vegetation due to similarities in forest characteristics such as height and stem density. Over 1 year, 0.6 ± 0.1 kg P ha−1 yr−1 were deposited in the open area, 0.8 ± 0.0 kg P ha−1 yr−1 beneath the successional vegetation, 0.5 ± 0.0 kg P ha−1 yr−1 beneath the artificial forest, and 1.9 ± 0.0 kg P ha−1 yr−1 beneath the mature forest. Results also showed an enrichment of P concentration beneath the artificial forest relative to the open area. Atmospheric P sources sustain 37% of the annual P demand in the mature forest, but only 13% in the successional vegetation. Thus, following deforestation, more of the P demand would have to be met from other sources that if unavailable, could lead to conditions where the forest does not recover.

Phosphorus (P) availability in soils can be a major factor limiting vegetation growth. Vegetation has adapted to this limitation by efficiently cycling P. Under conditions of limited P, microbes aid in mitigating P losses and enhancing P availability. Deforestation adversely impacts microbial populations. In turn, the loss of microbial biomass increases P losses and decreases P availability. This is important because P is not replenished biologically. Hence, when lost from the system, P only becomes available for uptake over geologic time scales. While vegetation clearly affects the microbial community and P cycling, it is not well-understood whether a loss of vegetation could lead to state changes in the vegetation dynamics. We develop a modeling framework and apply it to a P-limited cerrado ecosystem in Brazil with high intra-annual rainfall variability to examine whether deforestation can lead to a shift to a stable tree-less state. Following deforestation, we observed a decline in the microbial pool, a reduction in immobilization, and an increase in P losses. Although vegetation recovered from deforestation, a significant decline in the recalcitrant organic P pool occurred. A larger reduction in this pool led to a state change to the bare state from which the vegetation and microbial biomass did not recover. Results suggest that systems most susceptible to state changes may be those which have been previously deforested or those where the amount of P stored in the recalcitrant organic pool is low or becomes available very slowly.

Desertification is a change in soil properties, vegetation or climate, which results in a persistent loss of ecosystem services that are fundamental to sustaining life. Desertification affects large dryland areas around the world and is a major cause of stress in human societies. Here we review recent research on the drivers, feedbacks, and impacts of desertification. A multidisciplinary approach to understanding the drivers and feedbacks of global desertification is motivated by our increasing need to improve global food production and to sustainably manage ecosystems in the context of climate change. Classic desertification theories look at this process as a transition between stable states in bistable ecosystem dynamics. Climate change (i.e., aridification) and land use dynamics are the major drivers of an ecosystem shift to a “desertified” (or “degraded”) state. This shift is typically sustained by positive feedbacks, which stabilize the system in the new state. Desertification feedbacks may involve land degradation processes (e.g., nutrient loss or salinization), changes in rainfall regime resulting from land-atmosphere interactions (e.g., precipitation recycling, dust emissions), or changes in plant community composition (e.g., shrub encroachment, decrease in vegetation cover). We analyze each of these feedback mechanisms and discuss their possible enhancement by interactions with socio-economic drivers. Large scale effects of desertification include the emigration of “environmental refugees” displaced from degraded areas, climatic changes, and the alteration of global biogeochemical cycles resulting from the emission and long-range transport of fine mineral dust. Recent research has identified some possible early warning signs of desertification, which can be used as indicators of resilience loss and imminent shift to desert-like conditions. We conclude with a brief discussion on some desertification control strategies implemented in different regions around the world.

© 2016 Elsevier Ltd. Pressure on agricultural land has markedly increased since the start of the century, driven by demographic growth, changes in diet, increasing biofuel demand, and globalization. To better ensure access to adequate land and water resources, many investors and countries began leasing large areas of agricultural land in the global South, a phenomenon often termed "large-scale land acquisition" (LSLA). To date, this global land rush has resulted in the appropriation of 41million hectares and about 490km3of freshwater resources, affecting rural livelihoods and local environments. It remains unclear to what extent land and water acquisitions contribute to the emergence of water-stress conditions in acquired areas, and how these demands for water may be impacted by climate change. Here we analyze 18 African countries - 20Mha (or 80%) of LSLA for the continent - and estimate that under present climate 210km3year-1of water would be appropriated if all acquired areas were actively under production. We also find that consumptive use of irrigation water is disproportionately contributed by water-intensive biofuel crops. Using the IPCCA1B scenario, we find only small changes in green (-1.6%) and blue (+2.0%) water demand in targeted areas. With a 3°C temperature increase, crop yields are expected to decrease up to 20% with a consequent increase in the water footprint. When the effect of increasing atmospheric CO2concentrations is accounted for, crop yields increase by as much as 40% with a decrease in water footprint up to 29%. The relative importance of CO2fertilization and warming will therefore determine water appropriations and changes in water footprint under climate change scenarios.

The presence of forest vegetation can aid in preserving the permafrost layer by maintaining lower soil temperatures via the accretion of an organic layer at the surface. This layer has a low bulk density and low thermal conductivity (due also to high evapotranspiration rates by forest vegetation), which insulates ice-rich permafrost. Forest removal can lead to significant increases in the summer soil temperature, which increases the thickness of the active layer (i.e., the surficial layer above permafrost which thaws during summer and freezes again in winter) and causes a rise in the active layer towards the surface. When vegetation is sensitive to saturated soil conditions, a rise in the active layer can lead to the conversion of forested areas to wetlands. In this manuscript, we develop a modeling framework to relate vegetation–permafrost feedbacks to the emergence of multiple stable ecosystem states. Factors related to soil temperature and hydrologic characteristics of the system were examined to see how they affect the location of the stable and unstable states. This model was also used to examine how an increase in precipitation would affect the temporal dynamics of the active layer and vegetation. Results show that the presence of forest vegetation can enhance the resilience of the system in that it is less prone to state shifts following a disturbance. Understanding these dynamics is important given, (i) the rapid rate at which these systems can shift between states, (ii) the projected climatic changes for forested areas underlain by permafrost and (iii) the high rates of forest loss in these areas.

A process-based modeling framework is developed to examine the control that the soil moisture and soil carbon dynamics have on phosphorus (P) availability. The model accounts for soil moisture controls on several key fluxes that regulate P availability, including the decomposition rate of litter and humus, the mortality rate of microbes, the plant P demand and leaching of P from the rooting zone. To examine these dynamics over a wide range of time scales (i.e., from daily to seasonal to annual) we couple a zero-dimensional stochastic soil moisture model with a soil carbon model and a soil P model, all of which are run at the daily time scale. The soil carbon model accounts for the dynamics of the litter pool, humus pool, and microbial biomass pool, while the soil P model includes two inorganic P pools and four organic P pools that are coupled with the soil carbon pools. A mechanism is included in this framework to account for the contribution of enzymatic release to P availability, which has been shown to be an important source of P under conditions of low P availability. The model is applied to a Cerrado ecosystem located in Central Brazil where the phosphorus, carbon and hydrologic cycles have been well documented. Despite this application to a specific system, the modeling framework developed here is general and could be applied to any P-limited ecosystem for studies that seek to examine the controls on P-availability over time scales ranging from weeks to years. Results indicate that the model accurately captures the key driving variables controlling both the short-term (i.e., day to week) and long-term (i.e., year to decade) phosphorus, soil moisture and carbon dynamics. The effect of stochastic rainfall variability on P cycling is examined by running long-term simulations to obtain a probabilistic characterization of the state variables and fluxes under different rainfall regimes. The effect of a change in the precipitation regime suggests that such changes can significantly affect the dynamics of state variables and fluxes controlling P-availability. Moreover, results from the model illustrate the importance of the microbial retention of P, as a P cycling and conserving mechanism for P-limited ecosystems.

© 2015. American Geophysical Union. All Rights Reserved. Most of the human appropriation of freshwater resources is for agriculture. Water availability is a major constraint to mankind's ability to produce food. The notion of virtual water content (VWC), also known as crop water footprint, provides an effective tool to investigate the linkage between food and water resources as a function of climate, soil, and agricultural practices. The spatial variability in the virtual water content of crops is here explored, disentangling its dependency on climate and crop yields and assessing the sensitivity of VWC estimates to parameter variability and uncertainty. Here we calculate the virtual water content of four staple crops (i.e., wheat, rice, maize, and soybean) for the entire world developing a high-resolution (5 × 5 arc min) model, and we evaluate the VWC sensitivity to input parameters. We find that food production almost entirely depends on green water (>90%), but, when applied, irrigation makes crop production more water efficient, thus requiring less water. The spatial variability of the VWC is mostly controlled by the spatial patterns of crop yields with an average correlation coefficient of 0.83. The results of the sensitivity analysis show that wheat is most sensitive to the length of the growing period, rice to reference evapotranspiration, maize and soybean to the crop planting date. The VWC sensitivity varies not only among crops, but also across the harvested areas of the world, even at the subnational scale.

The analysis of the progression in time of scour depth at the contracted section of a bridge crossing is an important tool to the design and monitoring of bridge foundations. The nature of the riverbed material greatly affects the pace of scour development. In particular, in cohesive soils scour can develop at considerably long time scales and therefore a realistic scour depth analysis should refer to a long sequence of river flows. A method for estimating contraction scour in clear water and unsteady flow conditions is here proposed, which accounts for the cohesion of the bed material. The method is based on the use of the energy balance equation through the contraction. The dependence of scour rate on the bed material cohesion is expressed through a non-linear (empirically derived) scour erosion function. The method has been applied to estimate the scour depth in cohesive soils along different time spans. To this end, a probabilistic framework has been developed to fit a stochastic process to observed mean daily flow record and generate synthetic replicates of the historical series, which are used in a Monte Carlo procedure to calculate the probability density function and exceedance probability of the estimated scour depth values.

Damage to bridge crossings during flood events endangers the lives of the traveling public and causes costly disruptions to traffic flow. The most common causes of bridge collapse are scouring of the streambed and banks and erosion of highway embankments. This study couples a synthetic river flow simulation technique with a scour model for cohesive soils and determines the expected scour depth for a given lifetime of the bridge. A fractionally differenced autoregressive integrated moving average model generates synthetic streamflow sequences of the same length as the expected lifetime of the bridge. The scour model predicts the progression of scour depth through time in a multilayered soil. The model is used to determine the scour depth associated with different replicates of the synthetic flow sequences of the same length as the lifetime of the bridge. The probability distribution of scour depth is estimated by repeating this simulation procedure over a number of independent realizations of streamflow series for a given life of the bridge. This approach provides a framework for the probabilistic design and risk analysis of bridge foundations subjected to scour.

Even though the fundamental role played by nitrogen oxides in several important atmospheric processes is generally acknowledged, the understanding and quantitative evaluation of some mechanisms involved in the biogenic emission of N gases is still incomplete. Because of the number and variety of chemical, biological and physical processes involved, their mutual interaction and the strong nonlinearities, the dynamics of biogenic nitrogen oxide emission are complex and difficult to study. In this context, this study is centred on the link between soil water content and gaseous emissions of N2 and oxidized nitrogen gas, NO and N2O, from the ground. In particular, the study investigates how the probability distribution of soil moisture affects the distribution of the emission rate of nitrogen and nitrogen oxides when other limiting factors are absent. Some analytical tools are provided for the estimation of the steady-state probability density function of the emission rate. These tools are then used ...

The spatial and temporal dynamics of soil water content along a hillslope is the result of a number of complex and mutually interacting processes. This paper deals with the role of subsurface, unsaturated, lateral water flow and its links to climate, soil, and hillslope characteristics. The analysis focuses on the soil moisture dynamics at the daily time scale, averaged over the plant rooting depth, and accounts for the stochastic nature of precipitation as well as for the non-linear dependence of transpiration and hydraulic conductivity on soil moisture. The lateral fluxes of soil moisture are described by means of the one-dimensional Richards equation, and the probabilistic soil moisture dynamics is numerically investigated considering different conditions of climate, pedology, vegetation, and hillslope geometry. From the analysis two different regimes emerge: a humid one, characterized by an unsaturated lateral flow with significant spatial gradients of soil moisture along the hillslope, and a dry one, in which the topography does not affect the spatial distribution of the soil moisture. In the humid regime, the long-term average spatial pattern of soil water content is studied at different points along the hillslope using the mean, rms, and pdfs of soil moisture, as well as the components of the long-term water balance. All these analyses show how the soil moisture dynamics is the result of complex and non-local interactions between climate, soil, vegetation, and hillslope shape.

The nitrogen and carbon cycles in the broad-leafed savanna at Nylsvley (S. Africa) are modeled using the stochastic approach presented by Porporato et al. [Adv Water Res (this issue)]. An accurate representation of the hydrological mechanisms that control the nitrogen cycle at the daily time scale is shown to be necessary to capture the impact of the high-frequency variability of the soil moisture on the nitrogen and carbon dynamics. The fluctuations of the random precipitation forcing propagate to soil moisture, carbon, and nitrogen dynamics, giving rise to a gamut of fluctuations at different time scales. Long simulations are carried out to achieve a probabilistic characterization of the dynamics of the state variables under different rainfall regimes.

The influence of soil moisture dynamics on soil carbon and nitrogen cycles is analyzed by coupling an existing stochastic soil moisture model [Adv. Water. Resour. 24 (7) (2001) 707; Proc. R. Soc. Lond. A 455 (1999) 3789] to a system of eight nonlinear differential equations that describe the temporal evolution of the organic matter and the mineral nitrogen in the soil at the daily to seasonal time scales. Special attention is devoted to the modeling of the soil moisture control on mineralization and immobilization fluxes, leaching losses, and plant nitrogen uptake, as well as to the role played by the soil organic matter carbon-to-nitrogen ratio in determining mineralization and immobilization. The model allows a detailed analysis of the soil nitrogen cycle as driven by fluctuations in soil moisture at the daily time scale resulting from the stochastic rainfall variability. The complex ensuing dynamics are studied in detail in a companion paper [Adv. Water Resour. 26 (1) (2003) 59], which presents an application to the Nylsvley savanna in South Africa. The model accounts for the soil moisture control on different components of the nitrogen cycle on a wide range of time scales: from the high frequency variability of leaching and uptake due to the nitrate flushes after persistent rainfall following a period of drought, to the low frequency temporal dynamics of the soil organic matter pools. All the fluctuations in the various pools are statistically characterized in relation to their dependence on climate, soil, and vegetation characteristics.

The existence of an ice cover has important effects on the streamflow conditions as well as on the heat transfer between water bodies and the overlying atmosphere. This paper investigates the effects of climate variability on the termination of the ice season in the Baltic region. In particular, trends and fluctuations observed in the cryophenological records from this region are analyzed in detail, searching for possible connections with the North Atlantic Oscillation (NAO). The NAO seems to affect mostly the late-winter temperature (January–March) with a significant impact also on the mid-spring (April–May) period, when the air temperature is strongly correlated to the ice dates. A regional analysis shows the existence in the series of winter temperature (JFM) of the same fluctuations as the winter NAO. The same components can be found in the cryophenological records and (partly) in the series of spring temperature. Nevertheless, both ice phenology and spring temperature show the existence of a very well defined trend that is not detectable in the series of winter NAO at time scales of a century or longer. This leads to the conclusion that winter NAO has still a weak, though significant, effect on the regime of spring temperature in the Baltic region and explains the most significant fluctuating components embedded in the cryophenological records. However it is argued that other climatic forcings, related to CO 2 -induced regional and global warming, acting at the end of the ice season, are able to induce pronounced trends in the regime of spring temperature and have an important impact on the cryosphere leading to the earlier occurrence of ice break-up observed in the last several decades.

In the study of the evolutionary dynamics of soil moisture at a site, it is particularly important to define some characteristic properties of the temporal structure of the periods in which the soil water content is below certain levels indicative of water stress conditions in vegetation. The analysis of such properties provides an approach to establish some hydrologic basis for the understanding and modeling of ecosystems functioning in water-limited environments. This paper deals with a stochastic point process model of soil water balance. Expressions for both the mean number and the mean duration of time intervals during which the soil moisture is below a given threshold are analytically derived as a function of climate, soil, and vegetation. The seasonal mean value of water deficit is also analytically obtained. These properties are used to characterize the state of water stress in plants and to study its dependence on the interrelated dynamics. Estimates are included for the probability distributions of the frequency and duration of the stress and soil water deficit, for different hypotheses on climate, soil, and vegetation. Both the hydrologic and the ecologic implications of the results are briefly outlined.

The impact of climate fluctuations on the dynamics of soil moisture is studied through a stochastic model of soil water balance. The analysis focuses on the changes of soil water content induced by the interannual variability of rainfall observed at the decade-to-century timescale. Extensive data analyses have been performed to characterize the statistical properties of such a variability. Particular attention is paid to the year-to-year variability of the average value of soil moisture during the growing season because of its relevance to the mechanisms affecting the physiology of plants and the dynamics of ecosystems. It is found that the probability distribution of the average seasonal soil moisture may be either unimodal or bimodal depending on the different combinations of climate, soil, and vegetation parameters. The possible occurrence of a double mode has both hydrologic and ecologic implications that are analyzed here.

An important parameter in the assessment of the impact that conditions of limited soil water availability have on vegetation is the average length of the intervals during the growing season in which soil moisture is below some critical levels. These levels are related to the plant tolerance to water stress. The interannual rainfall variability induces important fluctuations on the average duration and frequency of the periods of water stress with important effects on the spatial and temporal structure of plant ecosystems. It is shown that the nonlinearities embedded in the dynamics controlling the soil water balance may drastically enhance the effects of the fluctuations present in the climatic forcing. Interannual climate variability leads to stronger year-to-year changes on the mean duration and frequency of periods of soil water deficit as well as to the emergence of preferential states in the probability distributions of these two variables. The sensitivity of these statistical properties is studied with respect to the characteristics of climate, soil, and vegetation.

The impact of climate fluctuations can be observed in the dynamics of vegetation and most particularly in the sensitive environment of savannas. In this paper we present a model for the local competition for soil moisture among neighboring vegetation. The initial condition for the model is a random field where at each point the soil moisture is the mean water content when there are no spatial interactions between sites. The mean soil moisture values account for stochasticity of climate and losses from evapotranspiration and leakage which depend on the existing water content. A spatial dynamics is then implemented based on the explicit minimization of the global water stress over the region. This approach explains the coexistence of herbaceous and woody plants in savannas as well as the changes in canopy density that have been documented in the southwest of the United States as a function of regional climatic fluctuations.

In the study of space-time rainfall it is particularly important to establish characteristic properties to guide both theoretical and modeling research efforts. In the present paper, new observational analyses on the scaling properties of time-evolving cumulated rainfall fields are presented, and a theoretical framework for their interpretation is introduced. It is found that the time evolution of the spatial organization of a cumulated rainfall field produces scaling relationships of spatial Variance versus time and characteristic values for the scaling exponent. The reproduction of these Values constitutes a basic requirement for spatial-temporal field generators in order to model important properties of real rainfall fields. It is then shown, on theoretical grounds, what properties the instantaneous rainfall intensity fields must obey in order to reproduce the experimental observations and how the size of the observation domain affects the scaling relationships. Some current stochastic models of space-time precipitation are finally discussed and analyzed in the light of the tools introduced, to show under what circumstances the models considered give acceptable results. Furthermore, it is shown that the assumption of an exponential time correlation function, used in many current rainfall models, is not compatible with observations.

[1] For some predominately phosphorus (P)-limited ecosystems, vegetation can be sustained under steady state conditions by atmospheric P inputs. The structure of the canopy influences deposition via the ability of the canopy to trap airborne P. This dependence suggests that a positive feedback may exist, which would have important impacts on the process of forest regeneration. One source of disturbance in the tropics is shifting cultivation. Over multiple cycles, studies have shown that shifting cultivation can lead to a large net loss of soil P, with a concomitant decline in forest biomass. In this study a model was developed to assess how shifting cultivation affects vegetation and phosphorus dynamics. This model is applied as a case study to a dry tropical forest system in the Southern Yucatan that is primarily P limited and has experienced shifting cultivation over several decades. Results show that following the second cycle, recovery would only be achieved after 100 years or longer in comparison to ∼30 years after one cycle. Examining the stable states of the system suggests that two stable states exist and that state changes as brought about by repeated disturbance can cause a shift to the other “low vegetation” stable state. Thus, for predominately P-limited ecosystems undergoing repeated disturbance, the depletion of soil P can significantly affect the long-term ability of the forest vegetation to recover. Results from this study have widespread implications, as 400 million ha of forest are affected by shifting cultivation.

[1] Shallow coastal bays provide habitat for diverse fish and invertebrate populations and are an important source of sediment for surrounding marshes. The sediment dynamics of these bays are strongly affected by seagrass meadows, which limit sediment resuspension, thereby providing a more favorable light environment for their own survival and growth. Due to this positive feedback between seagrass and light conditions, it has been suggested that bare sediment and seagrass meadows are potential alternate stable states of the benthos in shallow coastal bays. To investigate the stability and resilience of seagrass meadows subjected to variation in environmental conditions (e.g., light, temperature), a coupled model of vegetation–sediment–water flow interactions and vegetation growth was developed. The model was used to examine the effect of dynamically varying seasonal and interannual seagrass density on sediment resuspension, water column turbidity, and the subsequent light environment on hourly time steps and then run over decadal time scales. A daily growth model was designed to capture both belowground biomass and the growth and senescence of aboveground biomass structural components (e.g., leaves and stems). This allowed us to investigate how the annual and seasonal variability in shoot and leaf density within a meadow affects the strength of positive feedbacks between seagrass and their light environment. The model demonstrates both the emergence of bistable behavior from 1.6 to 1.8 m mean sea level due to the strength of the positive feedback, as well as the limited resilience of seagrass meadows within this bistable range.

[1] Forest canopies increase atmospheric inputs of nutrients to forest ecosystems by trapping dust and particulates. In tropical dry forests, this mechanism may contribute significant amounts of phosphorus, often a limiting nutrient in these ecosystems. Shifting cultivation may reduce the atmospheric inputs of phosphorus through the removal and restructuring of the forest canopy. We studied the impacts of land clearing by shifting cultivation in the southern Yucatan peninsula by measuring P in atmospheric bulk deposition and throughfall in two secondary forest stands and one mature forest stand to determine whether mature forests have greater throughfall P inputs than regenerating areas. From May to November 2007, we sampled rainfall in an open field and throughfall in three adjacent forest stands of different ages: 6 year old, 20 year old and mature (>60 year). We analyzed subsamples for inorganic and organic P. During the 7 month wet season, cumulative P input for the open field was 0.28 kg/ha, the 6 year stand accumulated 0.44 kg/ha, the 20 year stand accumulated 0.55 kg/ha and the mature stand accumulated 0.81 kg/ha. Organic P inputs were ∼50% of total P inputs in the open field, and 30–38% of total P inputs for the forest stands. The mature forest had significantly higher P concentrations and inputs than the open field or secondary forest stands, and forest stands had significantly greater P inputs than the open field, but there were no significant differences between the older and younger secondary forests. The repeated clearing of forests may thus reduce important P inputs to the ecosystem in the long-term.

[1] Changes in vegetation cover are known for their ability to modify the surface energy balance and near-surface microclimate conditions. A major change in vegetation composition that has been occurring in many dryland regions around the world is associated with the replacement of arid grasslands by desert shrublands. The impact of shrub encroachment on regional climate conditions remains poorly investigated, and, to date, it is unclear how this shift in plant community composition may affect the microclimate. Here we used concurrent meteorological observations at two adjacent sites dominated by Larrea tridentata shrubs and native grass species, respectively, in the northern Chihuahuan desert to investigate differences in nighttime air temperatures between the shrubland and grassland vegetation covers. The nighttime air temperature was found to be substantially higher (>2°C) in the shrubland than in the grassland, especially during calm winter nights. These differences in surface air temperature were accompanied by differences in longwave radiation and sensible and ground heat fluxes. We developed a one-dimensional model to show how longwave radiation emitted by the ground at night can explain the higher nighttime air temperature over the shrubland. Because of the larger fraction of bare soil typically existing in the shrub cover, the ground surface remains less insulated and more energy flows into the ground at the shrubland site than in the grassland during daytime. This energy is then released at night mainly as longwave radiation, which causes the differences in the nighttime air temperatures between the two land covers.

[1] Vegetation water stress plays an important role in the movement of water through the soil-plant-atmosphere continuum. However, the effects of water stress on evapotranspiration (ET) and other hydrological processes at the watershed scale remain poorly understood due in part to spatially and temporally heterogeneous conditions within the watershed, especially in areas of mountainous terrain. We used a spatially distributed model to understand and evaluate the relationship between water stress and ET in a forested mountain watershed during the snow-free growing season. Vegetation water stress increased as the growing season progressed, due to continued drying of soils, and persisted late into the growing season, even as vapor pressure deficit decreased with lower temperatures. As a result, ET became decoupled from vapor pressure deficit and became increasingly dependent on soil moisture later in the growing season, shifting from demand limitation to supply limitation. We found water stress and total growing season ET to be distributed nonuniformly across the watershed due to interactions between topography and vegetation. Areas having tall vegetation and low topographic index experienced the greatest water stress, yet they had some of the highest evapotranspiration rates in the watershed.

[1] When plants are both sensitive to salt levels in the root zone and able to modify the soil salt balance, changes in vegetation cover may affect the local hydrologic conditions and favor the accumulation of salt within different parts of the soil profile. In such cases a salt-vegetation feedback may exist, whereby both a state with vegetation cover, deep water table, and low salinity and a state with sparse or no vegetation, shallow water table, and high salinity can be stable. In this paper, we develop a modeling framework to relate vegetation–soil salinity feedbacks to the emergence of multiple stable states in the underlying dynamics. This model is used to simulate various scenarios involving changes in forcing parameters of salinity-vegetation dynamics using data from the Murray-Darling Basin. Results show the presence of a strong feedback resulting in bistable dynamics for a wide range of environmental conditions, which has the effect of reducing the resilience of plant ecosystems and the productivity of agricultural systems for areas where such a feedback can occur.

[1] Shallow coastal lagoons are environments where a dynamic equilibrium exists between water quality and seagrass cover. Dense seagrass canopies limit the resuspension of bed sediments thereby creating a clearer water column and a positive feedback for seagrass growth. Positive feedbacks are often associated with the existence of bistable dynamics in ecosystems. For example, a bare and a seagrass covered sediment bed could both be stable states of the system. This study describes a one-dimensional hydrodynamic model of vegetation-sediment-water flow interactions and uses it to investigate the strengths of positive feedbacks between seagrass cover, stabilization of bed sediments, turbidity of the water column, and the existence of a favorable light environment for seagrasses. The model is applied to Hog Island Bay, a shallow coastal lagoon on the eastern shore of Virginia. The effects of temperature, eutrophication, and bed grain size on bistability of seagrass ecosystems in the lagoon are explored. The results indicate that under typical conditions, seagrass is stable in water depths < 2.2 m (51% of the bay bottom deep enough for seagrass growth) and bistable conditions exist for depths of 2.2–3.6 m (23% of bay) where the preferred state depends on initial seagrass cover. The remaining 26% of the bay is too deep to sustain seagrass. Decreases in sediment size and increases in water temperature and degree of eutrophication shift the bistable range to shallower depths, with more of the bay bottom unable to sustain seagrass.

[1] In groundwater-dependent ecosystems, interactions between rainfall, water table fluctuations, and vegetation are exerted through the soil water content. The dynamics of soil moisture, in fact, are strongly coupled to fluctuations of the water table and, together, they control the overall ecosystem dynamics. We propose here a simple process-based stochastic model for the study of soil moisture dynamics at a generic depth, to complement the stochastic model of water table depth presented in the companion paper. The model presented here is based on a local and depth-dependent water balance driven by stochastic rainfall (marked Poisson noise) and accounting for processes such as rainfall infiltration, root water uptake, and capillary rise. We obtain a semianalytical formulation of the stationary probability distribution of soil water content at different depths, which is studied for different values of soil, climate, and vegetation parameters. The probability distributions are used to investigate the ecohydrology of groundwater-dependent ecosystems, including the quantitative description of the vegetation―water table―soil moisture interplay and the probabilistic analysis of root water uptake.

[1] Areas with a relatively shallow water table are environments where the groundwater plays a key role on the ecosystem function, and important interactions exist between hydrology and ecosystem processes. We propose here an analytical model to study the interactions between rainfall, water table, and vegetation in groundwater-dependent ecosystems. The water table dynamics are studied as a random process stochastically driven by a marked Poisson noise representing rainfall events. Infiltration, root water uptake, water flow to/from an external water body, and capillary rise are accounted for in a probabilistic description of water table fluctuations. We obtain analytical expressions for the steady state probability distribution of water table depth, which allows us to investigate the long-term behavior of water table dynamics, and their sensitivity to changes in climate, vegetation cover, and water management.

[1] A simplistic model of grass-shrub dynamics was used to investigate the role of grass demographic processes on grassland-shrubland dynamics when grasses are in competitive advantage over shrubs. The model suggests that a feedback between grass biomass and soil erosion may cause an abrupt transition to a shrubland state. The model explains how a simple change in either grass recruitment or grass mortality, presumably linked to climate change or grazing, could produce changes in Holocene flora and the conversion of grasslands to shrublands, which has been observed throughout the southwestern U.S. in the past 150 years.

[1] Dust aerosols can suppress rainfall by increasing the number of cloud condensation nuclei in warm clouds and affecting the surface radiation budget and boundary layer instability. The extent to which atmospheric dust may affect precipitation yields and the hydrologic cycle in semiarid regions remains poorly understood. We investigate the relationship between dust aerosols and rainfall in the West African Sahel where the dust-rainfall feedback has been speculated to contribute to sustained droughts. We find that the amount of dust loadings is negatively correlated with rainfall values, suggesting that dust entrained in the atmosphere can significantly inhibit rainfall in this region.

[1] This paper develops a geomorphological theory of tidal basin response (tidal instantaneous geomorphologic elementary response, or TIGER) to describe specific characteristics of tidal channel hydrodynamics. On the basis of the instantaneous unit hydrograph approach, this framework relates the hydrodynamics of tidal watersheds to the geomorphic structure of salt marshes and, specifically, to the distance traveled by water particles within the channel network and on the marsh surface. The possibility of determining the water fluxes from observations of geomorphic features is an appealing approach to the study of tidally driven flow rates. Our formulation paves the way to the application of recent results on the geomorphic structure of salt marshes and tidal networks to the determination of marsh creek hydrology. A case study shows how the asymmetry in the stage-velocity relation and the existence of velocity surges typical of the tidal hydrographs can be explained as an effect of the delay in the propagation of the tidal signal within the marsh area.

[1] The soil water content plays a fundamental role in a number of important environmental processes, including those involved in the water cycle, vegetation dynamics, soil biogeochemical cycles, and land-atmosphere interactions. Despite the recent efforts spent in the analytical modeling of the stochastic soil moisture dynamics in dryland ecosystems, the probabilistic characterization of the soil water balance in groundwater dependent environments is still missing, due to the complexity arising from the existence of shallow aquifers and capillary rise. This paper makes a first step in the direction of studying the stochastic soil water balance in groundwater dependent systems. An analytic probabilistic framework is developed for the modeling of soil moisture dynamics at the daily timescale for the case of bare soil conditions and in the presence of a shallow water table. This framework accounts for random rainfall occurrence, water table fluctuations, capillary rise, and soil evaporation. The stochastic soil water balance equation is analytically solved in the steady state, and the probability density functions of water table depth, soil moisture (as a function of depth), and evaporation are obtained and discussed for a variety of climate conditions and soil properties.

[1] Soil moisture is the environmental variable synthesizing the effect of climate, soil, and vegetation on the dynamics of water-limited ecosystems. 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 water is the limiting factor for vegetation in arid and semiarid ecosystems, a positive feedback could exist between soil moisture and woody 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 some field observations from savanna ecosystems located along the south-north rainfall gradient in the Kalahari, where the presence of relatively uniform sandy soils limits the effects of covarying factors. The data available from our field study suggest that the contrast between the soil moisture in the canopy and intercanopy space increases (with wetter soils under the canopy) with increasing levels of aridity. We hypothesize that this contrast may lead to a positive feedback and explore the implications of such a feedback in a minimalistic model. We found that when the feedback is relatively strong, the system may exhibit two stable states corresponding to conditions with and without tree canopy cover. In this case, even small changes in environmental variables may lead to rapid and largely irreversible shifts to a state with no tree canopy cover. Our data suggest that the tendency of the system to exhibit two (alternative) stable states becomes stronger in the more arid regions. Thus, at the desert margins, vegetation is more likely to be prone to discontinuous and abrupt state changes.

[1] Shrub encroachment, a widespread phenomenon in arid landscapes, creates “islands of fertility” in degraded systems as wind erosion removes nutrient-rich soil from intercanopy areas and deposits it in nearby shrub-vegetated patches. These islands of fertility generally are considered to be irreversible. Recently, fire has been observed to alter this pattern of resource heterogeneity through the redistribution of nutrients from the fertile islands of burnt shrubs to the surrounding bare soil areas. Despite the recognized relevance of both fires and wind erosion to the structure and function of arid ecosystems, the interactions between these two processes remains poorly understood. This study tests the hypothesis that fire-induced soil hydrophobicity developing in the soils beneath burned shrubs enhances soil erodibility by weakening the interparticle wet-bonding forces. To test this hypothesis, the effects of grass and shrub fires on changes in soil erodibility and on the intensity of fire-induced soil water repellency are compared at both the field and patch scales in heterogeneous arid landscapes. Higher water repellency was observed in conjunction with a stronger decrease in wind erosion threshold velocity around the shrubs than in grass-dominated patches affected by fire, while neither water repellency nor changes in threshold velocity was noticed in the bare soil interspaces. Thus, fires are found to induce soil hydrophobicity and to consequently enhance soil erodibility in shrub-vegetated islands of fertility. These processes create temporally dynamic islands of fertility and contribute to a decrease in resource heterogeneity in aridland ecosystems following fire.

[1] Soil water content is a key determinant of the health of terrestrial ecosystems. It plays a fundamental role in the feedbacks between the Earth and the atmosphere, as well as in all aspects of vegetation growth and composition. The dynamics of soil water in humid areas, and especially in wetlands, presents particularly challenging features for its quantitative description, since it needs to be linked to the intertwined stochastic fluctuations of the water table and the soil moisture of the unsaturated zone. These fluctuations are themselves dependent on the climate, soil, and vegetation of the region. The paper describes some of the most important problems that need to be considered in attempting to develop a quantitative framework for the ecohydrology of humid areas. The avenues of research suggested here will play a keystone role in the understanding of the complex dynamics of humid lands as well as in their scientifically based management in the face of a changing climate.

[1] In many terrestrial ecosystems, vegetation experiences limitation by different resources at different times. These resources include, among others, light, nutrients, and water. Frequently, however, leaf-level modeling frameworks that unite these limitations rely on empirical functions to scale stomatal conductance as a function of water stress. These functions use prescribed values of soil water content to mark the transition between water-stressed and unstressed conditions without accounting for the dependence of such a water content threshold on atmospheric and hydrologic conditions and nutrient availability. To address the phenomenon of a variable threshold to water stress, we combine an existing water-limited stomatal conductance model with an existing assimilation (photosynthesis)-limited stomatal conductance model. In this manner, we simulate variable controls on stomatal conductance and use a combination of the two models to define the threshold at which soil water content becomes limiting to transpiration. Modeled plant processes are used to define this water stress threshold as functionally dependent upon local environmental conditions (light, temperature, and atmospheric vapor pressure), parameters representing different vegetation types, and nutrient status. Simulations demonstrate that as environmental conditions become more favorable for assimilation, the likelihood of water stress increases. Specifically, there exist ranges of leaf temperature, light, and atmospheric humidity for which water stress is maximized.

[1] The high degree of spatial organization of dryland vegetation has been often explained invoking a number of different deterministic mechanisms without ever explicitly addressing the role of noise in the process of pattern formation. Noise is usually believed to act on ecosystems as a source of disorganized random fluctuations. However, noise is also known for its ability to induce ordered states in nonlinear systems. An alternative mechanism is here proposed, which explains vegetation pattern formation in mesic and subhumid savannas as the joint effect of fire randomness and fire-vegetation feedbacks. This mechanism is purely noise-induced and has no deterministic counterpart.

[1] Clarification of the coupled ecohydrological mechanisms that determine the spatial pattern and structural characteristics of vegetation in water-limited landscapes remains a vexing problem in both hydrological and vegetation sciences. A particular challenge is the fact that the spatial pattern of vegetation is both a cause and effect of variation in water availability in semiarid ecosystems. Here we develop a methodology to derive the landscape-scale distribution of water balance and soil moisture in a patchy vegetation mosaic based on the statistics of an underlying poisson distribution of individual tree canopies and their accompanying root systems. We consider the dynamics of water balance at a point to be dependent on the number of intersecting tree root systems and overlapping tree canopies. The coupling of individual pattern to landscape-scale distribution of soil water balance allows for investigations into the role of tree density, average canopy size, and the lateral extension of tree root systems on the spatiotemporal patterns of soil moisture dynamics, plant water uptake, and plant stress in a wide range of open woodland ecosystems. Our model is applied to southern African savannas, and we find that locations in the landscape that contain average vegetation structure correspond to conditions of minimum stress across a wide range of annual rainfall and vegetation densities. Furthermore, observed vegetation structural parameters are consistent with an optimization that simultaneously maximizes plant water uptake while minimizing plant water stress. Finally, the model predicts adaptive changes in the optimal lateral extent of plant roots which decreases with increasing rainfall along a regional gradient in mean annual precipitation.