Large-scale Catchment Research Articles

Evaluation of global continental hydrology as simulated by the Land-surface Processes and eXchanges Dynamic Global Vegetation Model

Abstract. Global freshwater resources are sensitive to changes in climate, land cover and population density and distribution. The Land-surface Processes and eXchanges Dynamic Global Vegetation Model is a recent development of the Lund-Potsdam-Jena model with improved representation of fire-vegetation interactions. It allows simultaneous consideration of the effects of changes in climate, CO2 concentration, natural vegetation and fire regime shifts on the continental hydrological cycle. Here the model is assessed for its ability to simulate large-scale spatial and temporal runoff patterns, in order to test its suitability for modelling future global water resources. Comparisons are made against observations of streamflow and a composite dataset of modelled and observed runoff (1986–1995) and are also evaluated against soil moisture data and the Palmer Drought Severity Index. The model captures the main features of the geographical distribution of global runoff, but tends to overestimate runoff in much of the Northern Hemisphere (where this can be somewhat accounted for by freshwater consumption and the unrealistic accumulation of the simulated winter snowpack in permafrost regions) and the southern tropics. Interannual variability is represented reasonably well at the large catchment scale, as are seasonal flow timings and monthly high and low flow events. Further improvements to the simulation of intra-annual runoff might be achieved via the addition of river flow routing. Overestimates of runoff in some basins could likely be corrected by the inclusion of transmission losses and direct-channel evaporation.

Hydrological impacts of rainwater harvesting (RWH) in a case study catchment: The Arvari River, Rajasthan, India. Part 1: Field-scale impacts

Rainwater harvesting (RWH), the small-scale collection and storage of runoff to augment groundwater stores, has been seen as a solution to the deepening groundwater crisis in India. However, hydrological impacts of RWH in India are not well understood, particularly at the larger catchment-scale. A key element to grasping RWH impact involves understanding the generated recharge variability in time and space, which is the result of variability in rainfall-runoff and efficiency of RWH structures. Yet there are very few reported empirical studies of the impact of RWH. Catchment-scale impacts are best studied using a water balance model, which would require a basic level of field data and understanding of the variability. This study reports the results of a 2-year field study in the 476 km2 semi-arid Arvari River catchment, where over 366 RWH structures have been built since 1985. Difficulties associated with working in semi-arid regions include data scarcity. Potential recharge estimates from seven RWH storages, across three different types and in six landscape positions, were calculated using the water balance method. These estimates were compared with recharge estimates from monitored water levels in 29 dug wells using the water table fluctuation method. The average daily potential recharge from RWH structures varied between 12 and 52 mm/day, while estimated actual recharge reaching the groundwater ranged from 3 to 7 mm/day. The large difference between recharge estimates could be explained through soil storage, local groundwater mounding beneath structures and a large lateral transmissivity in the aquifer. Overall, approximately 7% of rainfall is recharged by RWH in the catchment, which was similar in the comparatively wet and dry years of the field analysis. There were key differences between RWH structures, due to engineering design and location. These results indicate that recharge from RWH affects the local groundwater table, but also has potential to move laterally and impact surrounding areas. However, the greatest weakness in such analysis is the lack of information available on aquifer characteristics, in addition to geology and soil type.

Non‐native salmonids affect amphibian occupancy at multiple spatial scales

AbstractAim The introduction of non‐native species into aquatic environments has been linked with local extinctions and altered distributions of native species. We investigated the effect of non‐native salmonids on the occupancy of two native amphibians, the long‐toed salamander (Ambystoma macrodactylum) and Columbia spotted frog (Rana luteiventris), across three spatial scales: water bodies, small catchments and large catchments.Location Mountain lakes at ≥ 1500 m elevation were surveyed across the northern Rocky Mountains, USA.Methods We surveyed 2267 water bodies for amphibian occupancy (based on evidence of reproduction) and fish presence between 1986 and 2002 and modelled the probability of amphibian occupancy at each spatial scale in relation to habitat availability and quality and fish presence.Results After accounting for habitat features, we estimated that A. macrodactylum was 2.3 times more likely to breed in fishless water bodies than in water bodies with fish. Ambystoma macrodactylum also was more likely to occupy small catchments where none of the water bodies contained fish than in catchments where at least one water body contained fish. However, the probability of salamander occupancy in small catchments was also influenced by habitat availability (i.e. the number of water bodies within a catchment) and suitability of remaining fishless water bodies. We found no relationship between fish presence and salamander occupancy at the large‐catchment scale, probably because of increased habitat availability. In contrast to A. macrodactylum, we found no relationship between fish presence and R. luteiventris occupancy at any scale.Main conclusions Our results suggest that the negative effects of non‐native salmonids can extend beyond the boundaries of individual water bodies and increase A. macrodactylum extinction risk at landscape scales. We suspect that niche overlap between non‐native fish and A. macrodactylum at higher elevations in the northern Rocky Mountains may lead to extinction in catchments with limited suitable habitat.

Connection between streamwater mean residence time and bedrock groundwater recharge/discharge dynamics in weathered granite catchments

AbstractWe examined the connection between the mean residence time (MRT) of streamwater and the bedrock groundwater recharge/discharge dynamics in weathered granite catchments, the Kiryu Experimental Watershed (KEW), Japan. Several hydrological parameters including discharge, direct runoff ratio, baseflow recession, and bedrock infiltration were measured or estimated at subcatchments within KEW. Although large differences in the rainfall–runoff characteristics were observed among the small catchments, the differences became smaller and stabilized at the larger catchment scale. The key factor in the stabilization was the bedrock groundwater recharge and/or discharge dynamics, which reflected the differences in MRTs among the catchments. Our results suggest a dominant control of MRT by key geological factors, namely bedrock permeability and bedrock groundwater dynamics. In other words, MRT provides much information to consider when addressing hydrological scaling issues and/or estimating rainfall–runoff processes in a catchment, even an ungauged catchment. Copyright © 2010 John Wiley & Sons, Ltd.

Role of rural land use management in flood and coastal risk management

AbstractFlood risk is influenced by the rate and speed of rural land runoff within the catchment and by flow attenuation in the floodplain. Most research to date has demonstrated that the overall flood risk benefits of multiple small scale land management changes at a large catchment scale are difficult to determine and, in some situations, may not provide significant benefits in flood events. Measures may, however, offer potential for attenuation, improving flood warning times and therefore reducing flood damages. Climate change and major legislative drivers such as the Water Framework Directive will also further drive the need to consider different approaches to how land is managed and the impact that this may have on flood and coastal risk management.

Changes in the rates of floodplain and in-channel bench accretion in response to catchment disturbance, central Queensland, Australia

Analysis of the changes in rates of catchment sediment storage can provide material evidence of the impact of landscape disturbance on catchment sediment flux. A number of studies have suggested increased sediment yields from the rivers draining to the Great Barrier Reef since European settlement in the mid-nineteenth century. Many of these predictions, which indicate increases between four to ten times the pre-disturbance estimates, are based on large-scale catchment modelling that make some critical assumptions about pre-disturbance erosion rates and/or sediment delivery ratios. In addition, the majority have not been validated by empirical data. This study uses single-grain OSL dating and 137Cs depth profiles to determine pre- and post-floodplain accretion rates in Theresa Creek, a subcatchment of the dry-tropical Fitzroy River basin. We demonstrate that floodplain accretion rates have increased by three to four times since European settlement ( ca. A.D. 1850). Decreased rates of floodplain accretion since the mid-twentieth century at sites contributed to by gullied terrain suggest a decrease in the supply of sediment derived from gully networks. In contrast, floodplain accretion rates from areas dominated by cultivation remain high. Widespread in-channel benches deposited since European settlement are stable and appear to be important stores of large volumes of sediment. Low magnitude increases in post-disturbance floodplain sedimentation rates (3 to 4 times), in comparison to those reported from mainly temperate climates (10 to 100 times), are attributed to the naturally high sediment loads typical of dry-tropical catchments. Consequently, previous predictions of large post-disturbance increases in sediment yields from large dry-tropical catchments draining to the Great Barrier Reef are likely to be overestimates.

Fingerprinting upland sediment sources: particle size‐specific magnetic linkages between soils, lake sediments and suspended sediments

AbstractAccelerated erosion of fine‐grained sediment is an environmental problem of international dimensions. Erosion control strategies and targeting of mitigation measures require robust and quantitative identification of sediment sources. Here, we use magnetic ‘fingerprinting’ to characterize soils, and examine their affinity with and contribution to suspended sediments transported within two subcatchments feeding Bassenthwaite Lake, northwest England. A high‐resolution soil magnetic susceptibility survey was made using a field susceptometer (ZH Instruments, SM400 probe). Combining the spatial and vertical (down‐profile) soil magnetic data, a subset of soil profiles was selected for detailed, laboratory‐based magnetic remanence analyses. The magnetic properties of the catchment soils are highly particle size‐dependent. Magnetic analyses were performed on the 31–63 µm fraction, for particle size‐specific comparison both with the suspended sediments and lake sediments. Fuzzy cluster analysis groups the soil magnetic data into six clusters, apparently reflecting variations in parent material and horizon type, with three magnetically hard soils as unclassified outliers. Examination of the cluster affinity of the soils, suspended sediments and lake sediments indicates that topsoils of the upper Newlands Valley and subsoils around Keskadale Beck are a major source of the Newlands Beck suspended load, and the recent (post‐nineteenth century) sediments in the deep lake basin. Older lake sediments show strong affinity with a small number of the Derwent suspended sediments and one of the Glenderamackin soils. A large number of Derwent suspended sediments show no affinity with any of the soils or lake sediments, instead forming a coherent, discrete and statistically unclassified group, possibly resulting from mixing between the magnetically hard subsoils of the medium to high‐altitude Glenderamackin and Troutbeck areas and softer, lower altitude Glenderamackin soils. The lack of any affinity of these suspended sediments with the lake sediments may indicate deposition along the Derwent flood plain and/or in the shallow delta of Lake Bassenthwaite. Particle size‐specific magnetic fingerprinting is thus shown to be both highly discriminatory and quantitatively robust even within the homogeneous geological units of this catchment area. Such a methodological approach has important implications for small–large scale catchment management where sources of sediment arising from areas with uniform geology have been difficult to determine using other approaches, such as geochemical or radionuclide analyses. Copyright © 2009 John Wiley & Sons, Ltd.

Postwildfire hydrological response in an El Niño–Southern Oscillation–dominated environment

The rainfall‐runoff events following five fires that occurred within a 40‐year period in eucalypt forests of the Nattai catchment, southeastern Australia, were investigated to quantify the postwildfire hydrological response and to provide context for lower than expected erosion and sediment transport rates measured after wildfires in 2001. Daily rainfall and hourly instantaneous discharge records were used to examine rainfall‐runoff events in two gauged subcatchments (>100 km2) for up to 3 years after fire and compared with nonfire periods. Radar imagery, available from 2001, was used to determine the intensity and duration of rainfall events. Wildfires in the study catchment appear to have no detectable impact on surface runoff at the large catchment scale, regardless of fire severity, extent or time after fire. Instead, the magnitude of postfire runoff is related to the characteristics of rainfall after fire. Rainfall is highly variable in terms of annual totals and the number, size, and type of events. Rainfall events that cause substantial surface runoff are characterized by moderate‐high intensity falls lasting one or more days (≥1 year average recurrence interval). These are triggered by synoptic‐scale weather patterns, which do not reliably occur in the postfire window and are independent of broad‐scale climate dominated by the El Niño–Southern Oscillation (ENSO). This study highlights the importance of considering the characteristics of rainfall, as well as local factors, in interpreting the postfire hydrological response.

Which spatial discretization for distributed hydrological models? Proposition of a methodology and illustration for medium to large-scale catchments

Abstract. Distributed hydrological models are valuable tools to derive distributed estimation of water balance components or to study the impact of land-use or climate change on water resources and water quality. In these models, the choice of an appropriate spatial discretization is a crucial issue. It is obviously linked to the available data, their spatial resolution and the dominant hydrological processes. For a given catchment and a given data set, the "optimal" spatial discretization should be adapted to the modelling objectives, as the latter determine the dominant hydrological processes considered in the modelling. For small catchments, landscape heterogeneity can be represented explicitly, whereas for large catchments such fine representation is not feasible and simplification is needed. The question is thus: is it possible to design a flexible methodology to represent landscape heterogeneity efficiently, according to the problem to be solved? This methodology should allow a controlled and objective trade-off between available data, the scale of the dominant water cycle components and the modelling objectives. In this paper, we propose a general methodology for such catchment discretization. It is based on the use of nested discretizations. The first level of discretization is composed of the sub-catchments, organised by the river network topology. The sub-catchment variability can be described using a second level of discretizations, which is called hydro-landscape units. This level of discretization is only performed if it is consistent with the modelling objectives, the active hydrological processes and data availability. The hydro-landscapes take into account different geophysical factors such as topography, land-use, pedology, but also suitable hydrological discontinuities such as ditches, hedges, dams, etc. For numerical reasons these hydro-landscapes can be further subdivided into smaller elements that will constitute the modelling units (third level of discretization). The first part of the paper presents a review about catchment discretization in hydrological models from which we derived the principles of our general methodology. The second part of the paper focuses on the derivation of hydro-landscape units for medium to large scale catchments. For this sub-catchment discretization, we propose the use of principles borrowed from landscape classification. These principles are independent of the catchment size. They allow retaining suitable features required in the catchment description in order to fulfil a specific modelling objective. The method leads to unstructured and homogeneous areas within the sub-catchments, which can be used to derive modelling meshes. It avoids map smoothing by suppressing the smallest units, the role of which can be very important in hydrology, and provides a confidence map (the distance map) for the classification. The confidence map can be used for further uncertainty analysis of modelling results. The final discretization remains consistent with the resolution of input data and that of the source maps. The last part of the paper illustrates the method using available data for the upper Saône catchment in France. The interest of the method for an efficient representation of landscape heterogeneity is illustrated by a comparison with more traditional mapping approaches. Examples of possible models, which can be built on this spatial discretization, are finally given as perspectives for the work.

Linkage of sediment supply and transport processes in Miyagawa Dam catchment, Japan

Linkages between sediment supply by episodic geomorphic processes (obtained from aerial photographs and field surveys) and sediment transport (estimated from changes in sediment deposition behind a large dam) were investigated in the Miyagawa Dam catchment, central Japan. A total of 6667 landslides were confirmed in the period from 1965 to 2000 on the basis of seven temporal pairs of aerial photographs. Both the occurrence of landslides and discharge into the dam lake affect sediment yield, indicating that fluvial systems in Miyagawa Dam catchment are supply limited with respect to sediment. Sediment yields are affected by not only the initial failed volume of landslides but also the mobility of landslides and debris flows. In Miyagawa Dam catchment, percentages of landslides reaching channels varied from 56% in 1997–2000 to 75% in 1976–1981 and were correlated with maximum hourly rainfall. In addition, the mobility of debris flows was higher during periods with high maximum instantaneous discharge compared to lower discharge, suggesting that the water content both in initially failed materials and transported sediment controlled their mobility. Topography also affected the mobility of landslides/debris flows. For catchments >0.1 km2, the percentage of channel network length impacted by debris flows decreased with increasing catchment area due to reduced channel gradient. Thus both the magnitude of rainfall‐runoff events and the catchment topography affect how landslide sediment contributes to sediment yield at the large catchment scale.

Quantifying Fine-Sediment Sources in Primary and Selectively Logged Rainforest Catchments Using Geochemical Tracers

Detailed information on post-logging sediment dynamics in tropical catchments is required for modelling downstream impacts on communities and ecosystems. Sediment tracing methods, which are potentially useful in extending to the large catchment scale and longer time scales, are tested in primary and selectively logged rainforest catchments of Sabah, Borneo. Selected nutrient (P and N) and trace metal (Ni and Zn) concentrations are shown to discriminate surface, shallow subsurface and deep subsurface sediment sources. Analysis of channel-stored fine-sediment samples and use of an unmixing model allow the relative importance of these vertical sediment sources to be estimated and erosion processes to be inferred for catchments of contrasting size.

Modeling of Surface Runoff in Xitiaoxi Catchment, China

Xitiaoxi catchment is one of the most important catchments in the Taihu basin in China. Due to its significant contribution of surface runoff and associated nutrients input to Lake Taihu, understanding of the processes of surface runoff in this catchment is, therefore, of primary importance in quantifying water and nutrient balances for Lake Taihu. The generation of surface runoff in the catchment is mainly controlled by rainfall and land cover, so the variety of surface runoff in Xitiaoxi catchment is seasonal. Moreover, the annual change of surface runoff is distinct. Because of the diversity of land use and variety of hydrological characteristics, numerical simulation of the generation of surface runoff over this catchment is not straightforward. In this paper, attempts were made in applying the Large Scale Catchment Model (LASCAM) to Xitiaoxi catchment. The Xitiaoxi catchment is divided into 47 subcatchments connected via a river network based on topology. The model was first run in an optimization mode to calibrate the parameters against the observed runoff for the period of 1968-1977, and then was run in a prediction mode to try to reproduce the runoff for the next 10 years from 1978 to 1987. The model indicates that saturation excess runoff is probably the dominant process for the catchment. The modeling results indicate that water storage in shallow soils near the stream has a high correlation with daily rainfall, while the water storage changes in deep aquifers demonstrate an annual change trend, showing a rising level for wet seasons and a declining level for dry seasons. The success in modeling surface runoff leads to confidence in modeling nutrients transport as the next step of modeling work.

Modelling long-term mean hillslope erosion at a large catchment scale

Modelling long-term mean hillslope erosion at a large catchment scale

Mineralization of Organic Sulfur Delays Recovery from Anthropogenic Acidification

While SO4(2-) concentrations in runoff are decreasing in many catchments in Europe, present day S output still exceeds the S input for most forested catchments in Europe and North America. Here we report that a large part of the observed SO4(2-) in the runoff at a large-scale catchment study site (the Gårdsjön roof experiment in southwestern Sweden) originates from the organic S pool in the O horizon. Budget estimates comparing soil S pools showed reductions in the S pool of 57 mmol of S m(-2) in the O horizon and 26 mmol of SO4(2-) m(-2) in the mineral Bs horizon after excluding anthropogenic deposition for four years. There was an increase of about 1% per hundred in the delta34S(SO4), value of the mineral soil SO4(2-) between 1990 and 1995 (average and 95% confidence interval of 6.2 +/- 0.6 and 7.7 +/- 0.6% per hundred, respectively), but the delta34S(SO4) values in the E horizon are still much lower than the sprinkler water input of +19.7% per hundred, although the horizon has only a small extractable SO4(2-) pool. After nine years (1991-2000) of artificially supplying S inputs comparable with those amounts supplied by preindustrial rain, the amount of S in runoff still exceeded the input by 30%. This extra 30% corresponds to a loss of 3 mmol of S m(-2) year(-1), compared to the soil S organic O horizon pool of 1098 mmol m(-2) in 1990, suggesting that recovery is delayed for decades, at least.

Regional flow duration curve estimation in small ungauged catchments using instantaneous flow measurements and a censored data approach

Regional estimation of flow duration characteristics is important for water resources development at the small catchment scale. Regional analyses often fail to adequately represent the variability of the flow regime in small catchments (<50 km 2), especially in remote mountainous regions where the calibration data are sparse and refer to much larger catchment scales. This study suggests an approach in which regional data are combined with actual instantaneous flow data to construct a representative daily flow duration curve for small catchments. A regional dimensionless flow duration curve (FDC) is developed for a hydrologically homogeneous region in Western–Northwestern Greece and used to estimate the FDC in two small mountainous catchments within the region. A number of instantaneous flow measurements available at the two sites are used in a statistical representation of the flow regime from which an estimate of the mean annual flow at the sites is made, allowing the construction of the FDC from the regional curve. Results obtained are in good agreement with observed data and show significant estimation improvement over other methods commonly employed in the study region. A sensitivity analysis using Monte Carlo simulation is performed to establish reasonable sampling requirements for small ungauged catchments in the study area and similar mountainous areas in the Mediterranean region.

Modelling the effects of land-use modifications to control nutrient loads from an agricultural catchment in Western Australia

The estuary of the Swan and Canning Rivers in Western Australia is becoming increasingly prone to algal blooms, fish deaths and other biochemical problems that are thought to be associated with increasing eutrophication. Phosphorus and nitrogen enrichment are seen as the two most common causes of such eutrophication, with both elements being transported in streamflow and with concentrations strongly dependent upon land-use in the catchment. Many of the efforts to prevent and control eutrophication in the estuary are focused on managing land-use within the catchment. In this paper, the large-scale catchment model (LASCAM) is applied to Ellen Brook, a rural catchment located within the Swan River catchment, to simulate catchment exports of phosphorus and nitrogen, under a range of land cover scenarios that are designed to control the eutrophication. The scenarios, which are related to different management options for the catchment, are: (i) reforestation of agricultural land; (ii) reduction in fertiliser application; and (iii) urbanisation following a highway development. The model results show that: (i) full reforestation of agricultural land is expected to reduce phosphorus and nitrogen export by 50 and 85%, respectively; (ii) a proportionally greater reduction of phosphorus and nitrogen export occurs for smaller areas of reforestation than for larger areas; (iii) reduction in phosphorus fertiliser application produces a linear response with respect to phosphorus export; (iv) urbanisation increases runoff due to the larger impermeable areas causing an increase of overland flow during storms; and (v) phosphorus and nitrogen loads are expected to increase about 4 and 12%, respectively, during the 10 years following urbanisation.

Quantifying soil erosion by water in the UK: a review of monitoring and modelling approaches

The role of erosion by water in the UK is considered. A summary of available data describing water erosion is presented providing insights into rates of erosion from the hillslope scale to the large catchment scale. Evidence suggests that soil erosion rates in excess of acceptable thresholds occur on a wide range of soils and under a wide range of land uses throughout the country. Given the recent shift towards erosion modelling and away from erosion monitoring, discussion of the quality of existing available observed data in the context of model evaluation is made. Much quality data exist in the UK to describe erosion by water, but it is argued here that few datasets provide the necessary detail with which to evaluate model performance accurately, especially when the description of the spatial heterogeneity of soil loss is a goal. Furthermore, the paradox between data collection (to improve models) and erosion modelling (to replace data collection) is highlighted as an issue that must be addressed within the discipline if full use of datasets and improvement of models is to be made.

Convection-resolving precipitation forecasting and its predictability in Alpine river catchments

Predictability limitations in quantitative precipitation forecasting arising from small-scale uncertainties in the initial conditions are investigated for Alpine river catchments, with particular consideration of their implications on hydrological runoff forecasting. To this end, convection-resolving ensembles of limited-area simulations are performed using a nonhydrostatic numerical weather prediction (NWP) model, and results are analysed in terms of catchment-averaged precipitation. The applied ensemble strategy uses slightly modified initial conditions representing observational uncertainties, but identical lateral boundary conditions representing a perfectly predictable synoptic-scale forcing. A total of four case studies is carried out for different synoptic conditions leading to heavy precipitation. Ensemble integrations of 12 members are analysed for 24-h forecasting periods, with particular attention paid to precipitation in the Po basin and in its sub-catchments in the Lago Maggiore area. The simulations exhibit a large variability in the predictability of precipitation amounts, both from case to case and from catchment to catchment. It is demonstrated for an episode of thermal convection, that the predictability may be very low even in large-scale catchments of ∼50,000 km 2. In more synoptically dominated cases, predictability limitations appear to be restricted to catchments smaller than ∼10,000 km 2, while in one case predictability is found to be high in catchments as small as 200 km 2. Overall, the simulations show that precipitation forecasts for alpine river catchments may on occasions be critically affected by predictability limitations, even though the NWP model and the synoptic-scale forcing are assumed to be prefect. It is demonstrated that a substantial fraction of the predictability limitations is due to the scattered and unpredictable occurrence of convective cells, but the presence of convective precipitation alone does not necessarily limit predictability. It is also shown that the predictability is systematically higher in mountainous catchments.

Soluble reactive phosphorus levels in rainfall, cloud water, throughfall, stemflow, soil waters, stream waters and groundwaters for the Upper River Severn area, Plynlimon, mid Wales

Soluble reactive phosphorus (SRP) data are presented for rainfall, cloud water, soil waters, stream waters and groundwaters at the Plynlimon catchments in mid Wales to examine the hydrochemical functioning of inorganic phosphorus for an acidic and acid sensitive area characteristic of much of the UK uplands. In general, stream water concentrations are low compared to lowland areas. Average concentrations of SRP in rainfall and cloud water (0.3 and 0.9 μM l −1, respectively) are higher than in stream water with wider ranges in concentration (0–19.3 and 0–20.9 μM l −1, respectively). Throughfall and stemflow is enriched in SRP compared to rain and cloud water by a factor of approximately twofold and sixfold, respectively: the average concentrations and ranges are 0.73 and 0–6.61 μM l −1 for throughfall and 2.12 and 0–18.61 μM l −1 for stemflow. Soil water SRP concentrations measured in the surface layers of representative areas of podzol and gley soils, are further enriched with respect to inputs. Average concentrations and ranges for the L/F and Oh horizons in the podzols are 3.1 μM l −1 (range: 0.03–17.2 μM l −1) and 0.75 μM l −1 (range: 0.03–2.64 μM l −1), respectively. Correspondingly, the average values and ranges for the L/F and Oh horizons in the gley are 2 μM l −1 (range: 0.03–16.65 μM l −1) and 0.4 μM l −1 (range: 0.03–8.61 μM l −1). SRP concentrations in stream and ground water are lower than in atmospheric inputs and surface soil waters and show marked spatial variability. This variability is linked to three catchment features. (1) For streams draining podzolic soils, most of the SRP is retained by the catchment. For this situation, stream and ground waters have average concentrations of approximately 0.05 μM l −1 with a range of 0–1.47 μM l −1. There is no clear stream or groundwater SRP response to felling despite a large release of SRP from felling debris (brash) and the forest floor (L/F horizon) with average post-felling concentrations of 11.02 μM l −1 (0.40–155.0 μM l −1) and 23.60 μM l −1 (0.26–172.23 μM l −1), respectively. (2) For forested catchments with gley soils, stream water SRP concentrations are more variable with, in one case, much higher concentrations than for the podzol counterparts (range in average 0.05–0.46 μM l −1). (3) For the streams draining gley soils, felling results in a mixed SRP response. At the local scale (ditch drainage), there is a marked enrichment in SRP concentration (average concentrations increase from 0.05 to 1.31 μM l −1, with a peak concentration of 4.0 μM l −1). This response is consistent with the observed mobilisation of SRP from brash and forest floor material (post-felling mean concentrations of 9.39 and 11.94 μM l −1, respectively). However, stream water concentrations are an order of magnitude lower than observed in the soil waters implying considerable immobilisation of SRP between the soils and the stream. At the larger catchment scale, no discernable enrichment in SRP is observed following felling. The results are related to input–output budgets and the findings interpreted in terms of the dominant hydrogeochemical processes operative and environmental management issues.

Water resources and their use in food production systems

National scale assessments of global water scarcity predict that by 2050 as much as 66% of the world's future population will experience some water stress and that water requirements for food production will be a major issue. However, practical planning and management of water resources requires models with greater spatial and temporal resolution. Such models give improved (higher) estimates of the number of people at risk and how these might change as the population grows. Further improvements to these models require local/basin scale water resource data and, ultimately, the integration of relevant social, economic, institutional and legal information. An example of how progress might be made with this complex task of integration is given in the paper.¶ Although the above models are important for the identification and management of water resources issues, the crux of the food production issue lies in improving the efficiency with which food is produced from existing water and land resources. Water use efficiency1 can be increased in irrigated and rainfed systems using both physical and non-physical means. Substantial improvements are possible by maximizing transpiration and/or increasing the yield per unit of water transpired. The former is most practicable in systems with low water use efficiency and can be achieved mainly via managerial and/or institutional improvements, while the latter is more appropriate in systems with higher water use efficiency, via introducing agronomic and/or other technical improvements.¶There is also scope for improving water use efficiency and food production by considering larger catchment scales, where water can be recycled within irrigation schemes or moved between different rainfed parts of the landscape. This holistic approach to catchment water management allows the upstream/downstream impacts of improvements in agricultural water use efficiency to be taken into account and further emphasises the need for integration of both physical information and socio-economic factors if truly sustainable solutions are to be achieved. Future global change research, in both the physical and social sciences, has a major role to play in addressing this real and immediate challenge.¶¶ 1 Defined here as yield per unit area/water used to produce yield.