The Phosphorus Transfer Continuum: A Framework for Exploring Effects of Climate Change

Climate change effects on the P transfer continuum are examined. P export and eutrophication will likely increase under a changing climate. Climate change effects on catchment P processing are complex and poorly understood. Modeling of P flows and catchment P buffering is needed to inform management. Research should consider the effects of climate change on each tier of the continuum.

T he Earth's resources and integrity are under increasing pressure from a continually expanding population (Rockström et al., 2009). Agriculture is one such pressure that underpins society through the provision of food, yet it contributes to substantial alterations in our Earth system. Global temperatures might increase by at least 1.5°C by 2100, and extremes of climate, such as precipitation or temperature events, are frequently predicted in climate change scenarios (Christensen et al., 2013). Highly variable annual and seasonal shits in temperature and precipitation will have profound efects on agricultural function, output, and environmental impact. In particular, climate change will likely afect the patterns and eiciency of nutrient use in agricultural systems and subsequent luxes of these nutrients to the environment. hese changes are therefore critical to consider in terms of the future impact of food production systems on our environment, food security, and the resilience of food systems to climate change.
A critical nutrient for food production is phosphorus (P). Agriculture consumes vast quantities of P mined and puriied from phosphate rock (quadrupled since the mid-1900s), but these P inputs are ineiciently used in the food chain and have resulted in dramatic impairment of freshwater and marine ecosystems (Elser and Bennett, 2011). Signiicant eforts are being made to manage eutrophication impacts, but the environmental beneits of these eforts may become increasingly diicult to predict or control under a changing climate (e.g., bufer strip function) (Ockenden et al., 2017). Management adaptations that improve the environmental performance of agriculture in the short term consequently may not be adequate in the longer term under climate change (Kates et al., 2012). Standards and targets for eutrophication control may need to be reconsidered since the sensitivity of individual catchments to climate change is inherently dependent on catchment typology and its intrinsic P bufering properties, as well as the P pressure imposed by humans ( Fig. 1) (Doody et al., 2016).

Agricultural & Environmental Letters
Commentary Abstract: Phosphorus inputs to agriculture and their fate in the environment contribute to poor water quality and degradation of linked ecosystem services at great cost to society. Climate change is likely to alter the forms and timings of P luxes from land to water and their ecological impact, the efects of which are uncertain and need to be considered to inform future catchment management for eutrophication control. The P transfer continuum is an established conceptual model that we propose as a suitable framework to consider the potential efects of climate change on catchment P transfer. Consideration of this continuum suggests that predicted changes in temperature and precipitation will likely increase P transfer and associated eutrophication costs in some regions. Further research should examine climate change efects on each tier of the continuum to inform the necessary land management adaptations and transformations to ensure future food system P eiciency and resilience.

Core Ideas
• Climate change efects on the P transfer continuum are examined. • P export and eutrophication will likely increase under a changing climate. • Climate change efects on catchment P processing are complex and poorly understood. • Modeling of P lows and catchment P bufering is needed to inform management. • Research should consider the efects of climate change on each tier of the continuum.
he P transfer continuum, originally conceptualized by Haygarth et al. (2005), is a simple four-tiered model of source-mobilization-transport (or delivery)-impact that emphasizes the diferent scales and the interconnected dynamic nature of P mobility and mitigation in catchments (Haygarth et al., 2005;Withers and Haygarth, 2007) (Fig. 1). We propose this as a useful framework to improve conceptual understanding of the potential climate change efects on difuse P transfer across the land-water interface and the associated eutrophication risk. We highlight some of the complex efects of national and regional variability in predicted climate change scenarios on the continuum through its potential inluence on P inputs, P cycling, landscape hydrology, and eutrophication risk. A deeper understanding of these complex interactions is urgently needed to inform food system and catchment-based models to help synthesize net efects on terrestrial and aquatic ecosystems, to mitigate predicted risks to future food and water security, and to increase P eiciency.

The Phosphorus Transfer Continuum under Climate Change
Tier 1: Sources Sources of P include direct inputs that enter through the farm gate, such as fertilizers, imported animal feed stufs, and the application of imported livestock manure and other recycled bioresources to soil, which difer critically in amount, form, and timing depending on the agricultural system (Hale et al., 2015). Changing patterns of crop and animal production, and their yield potential due to climate change and agricultural intensiication, will alter decisions on inputs because P source is intrinsically linked to crop and animal demand. Climate change will have a large, and currently uncertain, inluence on regional land capability and suitability to grow speciic crop types for both human and animal consumption (Lobell and Gourdji, 2012;Rosenzweig et al., 2014) and therefore, P source inputs (Jobbágy and Sala, 2014). Increased frequency of extreme temperatures in some US states are likely to cause substantial declines (63-82% highest emissions scenario) in corn (Zea mays L.) and soybean [Glycine max (L.) Merr.] yields (Schlenker and Roberts, 2009). Conversely, increased temperatures in boreal and some temperate regions could raise agricultural output by extending or shiting growing seasons, allowing more crop choice and the possibility of double-cropping (Lobell and Gourdji, 2012), consequently altering regional P source input patterns and rates.
By 2050, meat and dairy consumption will increase significantly (Tilman and Clark (2014)). his will lead to changes in feed production and consumption, stocking densities, and the type of animals reared, afecting the amounts of livestock manure P that will need to be recycled back to the land. hese changes may also be constrained by alteration in land suitability due to climate change. Global trade and market demands will accordingly shit to relect altered production patterns and P inputs, some countries increasing their requirements and others reducing imports (Nesme et al., 2018). Coupled with increasing urbanization and diets that are more P-intensive (Metson et al., 2012), these infrastructural changes in production and trade can be expected to greatly alter P lows though the food system (rural and urban), afecting national, regional, and catchment P budgets and thus the potential for system P ineiciency and losses. Dynamic models are therefore required to analyze the potential inluence of a changing infrastructure on system P inputs, internal P lows (e.g., circulating and stored P), and P losses to ensure P use eiciency does not deteriorate under climate change (Withers et al., 2018).

Tier 2: Mobilization
he mobilization of P includes both solubilization and detachment of P from land, of which rainfall and temperature are critical regulators. Temperature-induced changes in soil C supply will alter both biotic and abiotic soil properties, afecting P dynamics, by increasing both the mineralization of organic P and the risk of soil particle dispersion by rain splash due to reduced soil cohesion (Macleod et al., 2012). Prolonged soil drying may induce P release via the oxidation of soil organic C -Fe and -Al associations but may also inhibit P mobilization by reducing P difusion rates from soil surfaces (Sheppard and Racz, 1984). In moist environments, higher ambient temperatures will increase P difusion and solubilization rates from soil and surface-applied soil amendments. Alternating wetting and drying will likely release soluble P through lysis of microbial cells, depending on microbial community structure and resilience to abiotic stress (Evans and Wallenstein, 2011).
Increased frequency and intensity of storm events under climate change will directly inluence the mobilization of soil particles and their P signatures, depending on soil type susceptibility to particle dispersion and detachment, the amount of bare soil present, and soil P status (Sharpley et al., 2008). In temperate regions, soils may also become anaerobic during prolonged periods of wet weather, causing P release when Fe-oxides are reduced (Scalenghe et al., 2012). Greater frequency and intensity of storm events will directly afect mobilization risk from organic and inorganic amendments when freshly applied to soil (incidental P loss), especially during winter (Liu et al., 2018). his is particularly important on livestock farms where manures are recycled to land frequently and therefore more susceptible to rainfall interception (Preedy et al., 2001), especially if best-practice application periods become reduced and storage capacity is limited. A greater fundamental understanding of the efect of climate change on these diferent mobilization processes is required for models to operate efectively.

Tier 3: Transport (Delivery)
Transport (alternatively called delivery) includes the delivery of P along pathways that occur via overland low, subsurface drainage, or leaching to groundwater. Climate change will alter pathway dominance and the speed of water routing, driven by hydrology and landscape topography, soil characteristics, and vegetation cover (Haygarth et al., 2005). For example, higher winter rainfall predicted in temperate regions may increase variable source areas of runof in close proximity to watercourses, while more intense storms may increase the extent of iniltration excess runof. In both cases, new critical source areas may become evident on farms where previously unseen runof zones coincide with high soil P, or where P sources are applied, requiring more sensitive land management, the success of which may be uncertain or costly (Renkenberger et al., 2017).
Greater intensity and duration of extreme precipitation events would generate more overland low from precipitation, leading to increased erosion, especially on hillslopes and where vegetative cover is low (Mullan, 2013). Landscape response to extreme precipitation is highly dependent on soil properties and indirectly dependent on land use and management (O'Neal et al., 2005). his places a greater emphasis on the importance of good soil management practices to optimize soil structure and iniltration capacity, such as intercropping and the development of climate extreme-resilient plant species. Conversely, regions that become drier and for longer may lose transport potential via erosion as runof ceases or frequency declines (Mullan et al., 2012). If soils are prone to extensive cracking, however, applied P can be rapidly transported down through the soil proile as preferential low ater rain (Simard et al., 2000). Prolonged dry spells also increase top-soil compaction; the hydrophobicity of soils consequently increases the routing of low in favor of surface runof, leading to greater P losses especially during intense rainfall events (Shakesby et al., 2000). Our understanding of hydrology-P interactions along surface and subsurface pathways and how these may be afected by climate change is poor, requiring innovative research in real-time high frequency monitoring.

Tier 4: Impact
Impact encompasses the ecological response of lotic and lentic waterbodies to P export from point and difuse sources and is regulated by the biological, chemical, and physical processes that govern retention (particulate P deposition, soluble P uptake by aquatic biota) and remobilization of these biotic and abiotic P stores (Jarvie et al., 2012). Under climate change, periods of low low or complete cessations of low that may occur in some regions will reduce the capacity for dilution, while increased ambient temperatures will accelerate nutrient cycling and biomass growth, consequently lowering dissolved oxygen concentrations, reducing the self-cleansing capacity of waterbodies and leading to increased eutrophication. Duan et al. (2012), for example, found that warming of streams entering the Chesapeake Bay consistently increased dissolved P luxes from bottom sediments to overlying water across diferent land uses.
he proliferation of algae may increase as waterbodies stay warmer for longer, potentially afecting macrophyte biomass, invertebrate and ish population health, and biodiversity as more tolerant species survive (Whitehead et al., 2009). Greater precipitation in some regions will increase river lows, increasing dilution capacity where difuse P inputs are low, reducing P concentrations. However, in some cases, >80% of total P load can be transported from headwater catchments during high-low events, resulting in high P loads, and poor water quality (Ockenden et al., 2016). Such high-low events can also move P-enriched sediments to lakes and coastal areas with increased risk of hypoxia during summer months. Increased nutrient loading to subarctic and temperate lakes, for example, have been responsible for increased harmful cyanobacterial blooms since 1800 (Taranu et al., 2015); hence, if high lows increase in frequency, the ecological impact of P could move increasingly from headwaters to lakes, coastal areas, and seas.

Conclusions and Research Needs
Phosphorus cycling and transfer through landscapes is highly complex, generating variable critical source areas and time lags across catchments, which are poorly understood in terms of their ecological impacts and management response. Climate change adds to this complexity because of uncertainties in climate predictions for different regions and because every component of this complex continuum will likely be affected by climate change, leading to variable P export patterns and ecological impacts across regions and catchments. Importantly, climate change will alter the ability of regions to grow specific crops or support high animal stocking densities, as well as the length of the growing season, affecting agricultural infrastructure. Phosphorus mobilization and transport under climate change will likely increase due to increased P solubilization and detachment and quicker delivery to waterbodies. Potentially new runoff zones and critical source areas will be created in some areas, which will require transformative management.
he sensitivity of individual catchments to climate change, however, is inherently dependent on catchment bufering capacity and the anthropogenic P pressure applied (Fig. 1). Consequently, research is needed to identify catchment resilience to climate change, and the rate of change, for each tier in the P transfer continuum. Catchment bufering properties should also be investigated to establish whether they may be enhanced by management adaptations (e.g., incremental changes to capture and recover P) and transformations (e.g., entire system transformation) to reduce P luxes and adverse ecological impacts and their associated costs. his commentary clearly identiies the need to improve modeling capability to predict climate change impacts on P processes and hydrology-P interactions in catchments, as well as the need to consider the wider impacts of agricultural infrastructure changes on P lows and eiciencies in the food system and the anthropogenic P pressures acting on catchments. In addition, we propose framing discussions on the potential implications of climate change on future P transfers by using the P transfer continuum, a simple model that can provide accessibility to knowledge for stakeholders and policymakers, as well as research direction for scientists.