The role of subsoil as a source or sink for phosphorus leaching.

The importance of subsoil features for phosphorus (P) leaching is frequently mentioned, but subsoil effects are still poorly documented. This study examined whether the subsoil of four agricultural Swedish soils (two sand and two clay) functioned as a source or sink for P leaching by measuring P leaching from intact soil columns with topsoil (1.05 m deep) and without topsoil (0.77 m deep) over 3 yr. One sandy soil with high topsoil P content (Olsen P, 84 mg kg) and high subsoil sorption capacity (P sorption index [PSI], 3.7 mmol kg) had low leaching of dissolved reactive P (DRP) from full-length and subsoil lysimeters (0.12 and 0.08 kg ha yr, respectively). The other sandy soil, with high Olsen P content in the topsoil and subsoil (27 and 19 mg kg, respectively) and low PSI in the subsoil (1.4 mmol kg), had high DRP leaching from full-length and subsoil lysimeters (3.33 and 3.29 kg ha yr, respectively). High P content at depth (Olsen P, 21 mg kg) in one clay soil resulted in relatively higher subsoil DRP contribution (89%) to total leaching than observed in the other clay soil (71%). These results indicate that the subsoil can act as source or sink for P leaching, depending on P content, degree of P saturation, and P sorption capacity, and therefore subsoil properties should be considered when selecting mitigation measures to reduce P leaching.

Turnover and losses of phosphorus in Swedish agricultural soils: long-term changes, leaching trends, and mitigation measures.

Transport of phosphorus (P) from agricultural fields to water bodies deteriorates water quality and causes eutrophication. To reduce P losses and optimize P use efficiency by crops, better knowledge is needed of P turnover in soil and the efficiency of best management practices (BMPs). In this review, we examined these issues using results from 10 Swedish long-term soil fertility trials and various studies on subsurface losses of P. The fertility trials are more than 50 years old and consist of two cropping systems with farmyard manure and mineral fertilizer. One major finding was that replacement of P removed by crops with fertilizer P was not sufficient to maintain soil P concentrations, determined with acid ammonium lactate extraction. The BMPs for reducing P leaching losses reviewed here included catch crops, constructed wetlands, structure liming of clay soils, and various manure application strategies. None of the eight catch crops tested reduced P leaching significantly, whereas total P loads were reduced by 36% by wetland installation, by 39 to 55% by structure liming (tested at two sites), and by 50% by incorporation of pig slurry into a clay soil instead of surface application. Trend analysis of P monitoring data since the 1980s for a number of small Swedish catchments in which various BMPs have been implemented showed no clear pattern, and both upward and downward trends were observed. However, other factors, such as weather conditions and soil type, have profound effects on P losses, which can mask the effects of BMPs.

Trade-offs between nitrogen and phosphorus removal with floodplain remediation in agricultural streams.

To improve water quality and reduce instream erosion, floodplain remediation along agricultural streams can provide multiple ecosystem services through biogeochemical and fluvial processes. During floodplain inundation, longer water residence time and periodic anoxic conditions can lead to increased nitrogen (N) removal through denitrification but also mobilization of phosphorus (P), impeding overall water quality improvements. To investigate the capacity for N and P processing in remediated streams, we measured potential denitrification and nitrous oxide production and yields together with potential P desorption and P fractions in floodplain and stream sediments in ten catchments in Sweden. Sediment P desorption was measured as equilibrium P concentration, using P isotherm incubations. Denitrification rates were measured with the acetylene inhibition method. Sediment nutrient process rates were combined with hydrochemical monitoring along remediated streams and their paired upstream control reaches of trapezoidal shape to determine the impact of floodplains on water quality. The correlation between floodplain denitrification rates and P desorption (r = 0.53, p = 0.02) revealed a trade-off between soluble reactive P (SRP) and nitrate removal, driven by stream water connectivity to floodplains. Nitrous oxide production was not affected by differences in P processing, but nitrous oxide yields decreased with higher denitrification and P desorption. The release of SRP from floodplains (0.03 ± 0.41 mg P kg-1 day-1) was significantly lower than from trapezoidal stream banks (0.38 ± 0.37 mg P kg-1 day-1), predicted by long-term SRP concentrations in stream water and floodplain inundation frequency. The overall impact of SRP release from floodplains on stream SRP concentrations in remediated reaches was limited. However, the remediated reaches showing increased stream SRP concentrations were also frequently inundated and had higher labile P content and coarse soil texture in floodplain sediments. To fully realize the potential for water quality improvements with constructed floodplains in agricultural streams, the promotion of denitrification through increased inundation should be balanced against the risk of P release from sediments, particularly in streams with high SRP inputs.

Toward better targeting of mitigation measures for reducing phosphorus losses from land to water: Andrew Sharpley's legacy in Norway and Sweden.

Nordic agriculture faces big challenges to reduce phosphorus (P) loss from land to water for improving surface water quality. While understanding the processes controlling P loss and seeking for P mitigation measures, Norwegian and Swedish researchers have substantially benefited from and been inspired by Dr. Andrew Sharpley's career-long, high-standard P research. Here, we demonstrate how Sharpley and his research have helped the Nordic researchers to understand the role of cover crops in cold environmental conditions, best manure P management practices, and ditch processes. His work on critical source area (CSA) identification and site assessment tool development have also greatly inspired our thinking on the targeting of mitigation measures and the contextualizing tools for Nordic climate, landscape, and soils. While reflecting on Sharpley's legacy, we identify several needs for Norwegian and Swedish P research and management. These include (1) tackling the challenges caused by local/regional unevenness in livestock density and related manure management and farm P surpluses, (2) identifying CSAs of P loss with high erosion risk and high P surplus, (3) obtaining more high-resolution mapping of soils with low P sorption capacity both in the topsoil and subsoil, (4) improving cross-scale understanding of processes and mitigation measures and proper follow-up of applied mitigation measures, and (5) increasing collaborations of researchers with farmers and farmers' advisory groups and watershed groups by developing high-quality educational courses and extension materials. The needs should be addressed in the context of the challenges and opportunities created by climate change.

Topsoil and subsoil properties influence phosphorus leaching from four agricultural soils.

Eutrophication, a major problem in many fresh and brackish waters, is largely caused by nonpoint-source pollution by P from agricultural soils. This lysimeter study examined the influence of P content, physical properties, and sorption characteristics in topsoil and subsoil on P leaching measured during 21 mo in 1-m-long, undisturbed soil columns of two clay and two sandy soils. Total P losses during the period varied between 0.65 and 7.40 kg ha. Dissolved reactive P was the dominant form in leachate from the sandy soils and one clay soil, varying from 48 to 76%. Particulate P dominated in leachate from the other clay soil, where low pH (5.2) in the subsoil decreased aggregate stability and thereby probably increased the dispersion of clay particles. Phosphorus leaching was small from soils with high P sorption index (PSI) and low P saturation (<10% of PSI) in the subsoil, even though extractable P (Olsen P) in the topsoil was high, and large from a soil with low sorption capacity and high P saturation (>35% of PSI) in the profile. High sorption capacity in the subsoil was more important for P leaching in sandy soils than in clay soils with macropore flow, where the effect of high sorption capacity was reduced due to less interaction between percolating water and the soil matrix. The results suggest that P leaching is greatly affected by subsoil properties and that topsoil studies, which dominate current research, are insufficient for assessing P leaching in many soils.

Barium as a potential indicator of phosphorus in agricultural runoff.

In many catchments, anthropogenic input of contaminants, and in particular phosphorus (P), into surface water is a mixture of agricultural and sewage runoff. Knowledge about the relative contribution from each of these sources is vital for mitigation of major environmental problems such as eutrophication. In this study, we investigated whether the distribution of trace elements in surface waters can be used to trace the contamination source. Water from three groups of streams was investigated: streams influenced only by agricultural runoff, streams influenced mainly by sewage runoff, and reference streams. Samples were collected at different flow regimes and times of year and analyzed for 62 elements using ICP-MS. Our results show that there are significant differences between the anthropogenic sources affecting the streams in terms of total element composition and individual elements, indicating that the method has the potential to trace anthropogenic impact on surface waters. The elements that show significant differences between sources are strontium (p < 0.001), calcium (p < 0.004), potassium (p < 0.001), magnesium (p < 0.001), boron (p < 0.001), rhodium (p = 0.001), and barium (p < 0.001). According to this study, barium shows the greatest potential as a tracer for an individual source of anthropogenic input to surface waters. We observed a strong relationship between barium and total P in the investigated samples (R(2) = 0.78), which could potentially be used to apportion anthropogenic sources of P and thereby facilitate targeting of mitigation practices.

Land use, geology and soil properties control nutrient concentrations in headwater streams.

Nutrient losses from headwater catchments (<50 km2) cause eutrophication problems downstream. Catchment properties are strongly reflected in the levels of nutrient concentrations in headwater streams. Based on measurements of total and dissolved nitrogen (TN, DN) and phosphorus (TP, DP) in 235 small headwater streams, we showed that proportion of arable land in a catchment had the strongest positive effect on nutrient concentrations, with coefficient of determination (R2) of 0.54, 0.64, 0.45, and 0.51 for TN, DN, TP, and DP, respectively. In contrast, increased proportion of forest and wetland led to lower nutrient concentrations in streams. The geological composition of catchments had a major influence on the soil properties. In turn, certain soil properties, such as clay content and content of aluminum (Al), an important binding agent of P, influenced losses of particulate P (PP) and DP, respectively. Consequently, by using soil properties as a link between geology and water quality, areas potentially sensitive to nutrient losses were identified by classifying bedrock categories into three geological groups. Approximately 25% of Swedish arable land was identified as potentially sensitive. Sensitive catchments were found in regions with sedimentary bedrock and showed higher concentrations of dissolved nutrient fractions even when the proportion of agricultural land was small, indicating higher background concentrations.

Variability in fluvial suspended and streambed sediment phosphorus fractions among small agricultural streams.

Agriculture is a major source of sediment and particulate phosphorus (P) inputs to freshwaters. Distinguishing between P fractions in sediment can aid in understanding its eutrophication risk. Although streams and rivers are important parts of the P cycle in agricultural catchments, streambed sediment and especially fluvial suspended sediment (FSS) and its P fractions are less studied. To address this knowledge gap, seasonal variations in FSS P fractions and their relation to water quality and streambed sediment were examined in three Swedish agricultural headwater catchments over 2 yr. Sequential fractionation was used to characterize P fractions in both streambed sediment and FSS. All catchments had similar annual P losses (0.4-0.8 kg ha-1 ), suspended solids (124-183 mg L-1 ), and FSS total P concentrations (1.15-1.19 mg g-1 ). However, distribution of P fractions and the dominant P fractions in FSS differed among catchments (p < .05), which was most likely dependent on differences in catchment geology, clay content, external P sources, and flow conditions. The most prominent seasonal pattern in all catchments was found for iron-bound P, with high concentrations during low summer flows and low concentrations during winter high flows. Streambed sediment P fractions were in the same concentration ranges as in FSS, and the distribution of the fractions differed between catchments. This study highlights the need to quantify P fractions, not just total P in FSS, to obtain a more complete understanding of the eutrophication risk posed by agricultural sediment losses.

How effective are River Basin Management Plans in reaching the nutrient load reduction targets?

Riverine nutrient loads are among the major causes of eutrophication of the Baltic Sea. This study applied the Soil & Water Assessment Tool (SWAT) in three catchments flowing to the Baltic Sea, namely Vantaanjoki (Finland), Fyrisån (Sweden), and Slupia (Poland), to simulate the effectiveness of nutrient control measures included in the EU's Water Framework Directive River Basin Management Plans (RBMPs). Moreover, we identified similar, coastal, middle-sized catchments to which conclusions from this study could be applicable. The first modelling scenario based on extrapolation of the existing trends affected the modelled nutrient loads by less than 5%. In the second scenario, measures included in RBMPs showed variable effectiveness, ranging from negligible for Slupia to 28% total P load reduction in Vantaanjoki. Adding spatially targeted measures to RBMPs (third scenario) would considerably improve their effectiveness in all three catchments for both total N and P, suggesting a need to adopt targeting more widely in the Baltic Sea countries.

Process based modelling of phosphorus losses from arable land.

Improved understanding of temporal and spatial Phosphorus (P) discharge variations is needed for improved modelling and prioritisation of abatement strategies that take into account local conditions. This study is aimed at developing modelling of agricultural Phosphorus losses with improved spatial and temporal resolution, and to compare the accuracy of a detailed process-based model with a rainfall-runoff coefficient-based model. The process-based SWAT model (Soil and Water Assessment Tool) was implemented for five river basins in central Sweden, and results compared with the rainfall-runoff coefficient-based model WATSHMAN (Watershed Management System) for one of these river basins. Parameter settings and attribute values were adapted to Scandinavian soil conditions, crops and management practices. Model performance regarding flow dynamics was overall satisfactory. Comparable results were achieved at several scales. The modelled P load was of high accuracy for the days when monitoring data were available for validation, generally once a month. Modelled monthly P load did not fit as well with averaged monthly monitoring load values, mainly since monthly monitoring often partly or entirely misses the peak flows. The comparison of SWAT and WATSHMAN gave slightly better results for the process-based model (SWAT). Better spatial resolution for input data such as Soil-P content and agricultural management practices will be required to reach modelling results that enable identification of measures adapted to local conditions.

Optimizing placement of constructed wetlands at landscape scale in order to reduce phosphorus losses.

Constructed wetlands (CWs) are one of the main countermeasures to reduce diffuse phosphorus (P) losses, but there is still a lack of systematic guidance accounting for spatially variable effects of hydraulic and P load on P retention. We present a three-step modelling approach for determining suitable placement of CWs in four different size groups (0.1-1.0 ha), based on incoming hydraulic and P load. The modelled hypothetical CW area was only 17% of that previously estimated and area of efficient CWs is even lower. The mean area-specific P retention increased with CW size. However, the spatial variation in retention was large for all size groups and largest (6-155 kg ha-1 year-1) for the smallest CWs due to highly variable incoming P loads, showing the possible benefits of targeted placement of CWs. The presented modelling approach has also flexibility to include and account for possible future changes in land cover and management.

Particulate phosphorus and suspended solids losses from small agricultural catchments: Links to stream and catchment characteristics.

Excessive phosphorus (P) inputs from agriculture are well established as a contributor to freshwater eutrophication. Decreasing these inputs is an important step in improving the ecological state of impaired waters. Particulate P (PP) is a significant contributor to diffuse P inputs in agricultural catchments. Identifying the main correlates for PP losses is an important step in reducing these inputs. However, there are few studies of long term temporal and spatial dynamics of PP in agricultural streams. Here, we investigate the relative importance of hydrology, catchment characteristics and geochemistry on PP concentrations and fluxes in agricultural headwaters. We evaluate long-term monitoring data from eleven small (<35 km2) Swedish catchments with at least seven years of measured flow and flow proportional water quality sampling. Using parametric and non-parametric regression together with principal components analysis (PCA), we identify in-stream and catchment variables relevant for predicting PP concentrations, e.g., suspended solids concentrations (SS), soil texture and average catchment soil P content, measured as ammonium lactate/acetic acid extractable P (P-AL). We show that PP is primarily correlated to SS concentrations, which in turn are correlated to average clay content and land use. However, the SS:PP relationships differ between catchments. No correlation between PP concentrations in the stream and soil P content was found. An increasing clay content decreases the slope of the relationship between SS and PP, i.e., in catchments with higher clay content, less PP is transported per unit SS. The PP/SS ratio increased significantly (p < 0.05) over time in four catchments, despite limited changes in SS or PP concentrations. Our study highlights the importance of long time series since the enrichment of P on SS in the streams is only detected when using long term monitoring data.

From single fields to river basins: Identification of critical source areas for erosion and phosphorus losses at high resolution.

Concentrations of phosphorus (P), the main limiting nutrient in freshwater ecosystems, need to be reduced, but this is difficult due to high spatial and temporal variations and limited resources. Reliable targeting of critical source areas, such as erosion-prone fields and parts of fields, is necessary to improve the cost efficiency of mitigation measures. We used high-resolution (2 m × 2 m) distributed modelling to calculate erosion risk for a large area (202 279 km2) covering > 90% of Swedish arable land. Comparison of model results with independent farmers' observations in a pilot catchment showed high spatial agreement. The modelled worst case scenario produced reasonable quantitative results comparable to measured 90th percentile values of suspended sediment (SS) loads at both field and small catchment scale (R2 = 0.81, p < 0.001). Overall, loads of SS, especially during extreme episodes, strongly governed losses of unreactive P and total P at both field and catchment scale.

Hydrologic Extremes and Legacy Sources Can Override Efforts to Mitigate Nutrient and Sediment Losses at the Catchment Scale.

Combating eutrophication requires changes in land and water management in agricultural catchments and implementation of mitigation measures to reduce phosphorus (P), nitrogen (N) and suspended sediment (SS) losses. To date, such mitigation measures have been built in many agricultural catchments, but there is a lack of studies evaluating their effectiveness. Here we evaluated the effectiveness of mitigation measures in a clay soil-dominated headwater catchment by combining the evaluation of long-term and high-frequency data with punctual measurements upstream and downstream of three mitigation measures: lime-filter drains, a two-stage ditch, and a sedimentation pond. Long-term hydrochemical data at the catchment outlet showed a significant decrease in P (-15%) and SS (-28%) and an increase in nitrate nitrogen (NO-N, +13%) concentrations. High-frequency (hourly) measurements with a wet-chemistry analyzer (total and reactive P) and optical sensor (NO-N and SS) showed that the catchment is an abundant source of nutrients and sediments and that their transport is exacerbated by prolonged drought and resuspension of stream sediments during storm events. Lime-filter drains showed a decrease in SS by 76% and total P by 80% and an increase in NO-N by 45% compared with traditional drains, potentially indicating pollution swapping. The effectiveness of two-stage ditch and sedimentation pond was less evident and depended on the prevalent hydrometeorological conditions that drove the resuspension of bed sediments and associated sediment-bound P transport. These results suggest that increased frequency of prolonged drought due to changing weather patterns and resuspension of SS and sediment-bound P during storm events can override the generally positive effect of mitigation measures.

A Global Perspective on Phosphorus Management Decision Support in Agriculture: Lessons Learned and Future Directions.

The evolution of phosphorus (P) management decision support tools (DSTs) and systems (DSS), in support of food and environmental security has been most strongly affected in developed regions by national strategies (i) to optimize levels of plant available P in agricultural soils, and (ii) to mitigate P runoff to water bodies. In the United States, Western Europe, and New Zealand, combinations of regulatory and voluntary strategies, sometimes backed by economic incentives, have often been driven by reactive legislation to protect water bodies. Farmer-specific DSSs, either based on modeling of P transfer source and transport mechanisms, or when coupled with farm-specific information or local knowledge, have typically guided best practices, education, and implementation, yet applying DSSs in data poor catchments and/or where user adoption is poor hampers the effectiveness of these systems. Recent developments focused on integrated digital mapping of hydrologically sensitive areas and critical source areas, sometimes using real-time data and weather forecasting, have rapidly advanced runoff modeling and education. Advances in technology related to monitoring, imaging, sensors, remote sensing, and analytical instrumentation will facilitate the development of DSSs that can predict heterogeneity over wider geographical areas. However, significant challenges remain in developing DSSs that incorporate "big data" in a format that is acceptable to users, and that adequately accounts for catchment variability, farming systems, and farmer behavior. Future efforts will undoubtedly focus on improving efficiency and conserving phosphate rock reserves in the face of future scarcity or prohibitive cost. Most importantly, the principles reviewed here are critical for sustainable agriculture.

Targeting critical source areas for phosphorus losses: Evaluation with soil testing, farmers' assessment and modelling.

Diffuse phosphorus (P) losses from arable land need to be reduced in a cost-efficient way, taking into account their temporal and spatial variability. This study, based on 16 farms across southern Sweden, examined possibilities for identifying critical source areas for P losses based on the combined results of high-resolution erosion modelling, independent risk assessments by farmers, soil survey and SWOT analysis performed by farmers. Statistically significant differences in dissolved P release were found between soil P test classes in the studied area, whereas soil textural classes and not P content governed potential mobilisation of soil particles and unreactive P. Spatial comparison of problem areas identified by farmers and modelled features showed that the modelled erosion pathways intersected 109 in a total of 128 (85%) observed problem areas. The study demonstrates the value in involving farmers in the identification of critical source areas in order to select and support implementation of effective countermeasures.

Conditional phosphorus index as an educational tool for risk assessment and phosphorus management.

Phosphorus index (PI) is a risk-assessment tool that combines phosphorus (P) source factors and transport factors to rank the vulnerability of fields to P losses. Here we present the structure and concepts of conditional PI, developed as an educational and P-management tool adjusted for Swedish conditions. Because the significance of certain factors for P losses depends on their interplay with other factors, conditional rules are needed for a more accurate process description and quantification. Accounting for P losses through the soil profile, separate calculations for reactive and unreactive P and a changed approach to P loss assessment from erosion losses are some of the new features included in the tool presented here. The performance of the tool was tested by comparing the calculated PI values with measured annual P transport from seven observation fields included in a Swedish water quality monitoring program. This first test indicated that the tool could be used successfully for P loss risk assessment.

Distributed, high-resolution modelling of critical source areas for erosion and phosphorus losses.

Phosphorus losses from arable land need to be reduced to prevent eutrophication of surrounding waters. Owing to the high spatial variability of P losses, cost-effective countermeasures need to target parts of the catchment that are most susceptible to P losses. Field surveys identified critical source areas for overland flow and erosion amounting to only 0.4-2.6 % of total arable land in four different catchments in southern Sweden. Distributed modelling using high-resolution digital elevation data identified 72-96 % of these observed erosion and overland flow features. The modelling results were also successfully used to predict occurrence of overland flow and rill and gully erosion in a catchment in central Sweden. Such exact high-resolution modelling allows for accurate placement of planned countermeasures. However, current legislative and environmental subsidy programmes need to change their approach from income-loss compensation to rewarding high cost effectiveness of implemented countermeasures.

Screening risk areas for sediment and phosphorus losses to improve placement of mitigation measures.

Identification of vulnerable arable areas to phosphorus (P) losses is needed to effectively implement mitigation measures. Indicators for source (soil test P, STP), potential mobilization by erosion (soil dispersion), and transport (unit-stream power length-slope, LS) risks were used to screen the vulnerability to suspended solids (SS) and P losses in two contrasting catchments regarding topography, soil textural distribution, and STP. Soils in the first catchment ranged from loamy sand to clay loam, while clay soils were dominant in the second catchment. Long-term SS and total P losses were higher in the second catchment in spite of significantly lower topsoil STP. A higher proportion of areas in the second catchment were identified with higher risk due to the significantly higher risk of overland flow generation (LS) and a significantly higher mobilization risk in the soil dispersion laboratory tests. A simple screening method was presented to improve the placement of mitigation measures.

Future agriculture with minimized phosphorus losses to waters: Research needs and direction.

The series of papers in this issue of AMBIO represent technical presentations made at the 7th International Phosphorus Workshop (IPW7), held in September, 2013 in Uppsala, Sweden. At that meeting, the 150 delegates were involved in round table discussions on major, predetermined themes facing the management of agricultural phosphorus (P) for optimum production goals with minimal water quality impairment. The six themes were (1) P management in a changing world; (2) transport pathways of P from soil to water; (3) monitoring, modeling, and communication; (4) importance of manure and agricultural production systems for P management; (5) identification of appropriate mitigation measures for reduction of P loss; and (6) implementation of mitigation strategies to reduce P loss. This paper details the major challenges and research needs that were identified for each theme and identifies a future roadmap for catchment management that cost-effectively minimizes P loss from agricultural activities.