Three-dimensional perspectives on revisiting the legacy impacts of non-point source phosphorus (2024)

Phosphorus (P) is a vital biogenic element required by Earth's ecosystems. Water and food security heavily depend on the sustainable supply and management of P [1]. However, since the 20th century, human activities (e.g. industrialization, urbanization, and agricultural intensification) have significantly disturbed P cycling within terrestrial ecosystems where anthropogenic P has precipitated a near doubling of global P loss from lands to surface freshwaters, increasing from 5 to 9 Tg P a⁻¹ [2]. Such a significant increase has emerged as a major driver behind the concern of excess P concentrations and loads in freshwaters that aggravate eutrophication [3]. Management of P loss has become a global challenge in agricultural sustainability and improvement of aquatic environments [4].

Numerous studies underscored the predominant contribution of non-point sources (NPSs) to P loss through runoff and soil erosion. Managerial and structural practices to control NPS P pollution have been widely implemented globally over the past few decades. However, long-term water quality monitoring results from many critical water bodies (e.g. the Chesapeake Bay, Lake Erie, and Mississippi River) indicated that these practices failed to deliver expected improvements in water quality in terms of P within predicted timescales [5]. It has become increasingly evident that the NPS P accumulated in the environment that serves as a chronic source to freshwaters is the root of the 'wicked problem'. Especially in agroecosystem, fertilizers and manures are often massively applied to farmlands to ensure regional food supply, resulting in soil P content far exceeding the demands of crop production. The proportion of P utilized is just under 30%, whereas the remainder is less susceptible to be used by future crops [6]. The unutilized P is partially accumulated and retained in the soil, a portion leaches into the groundwater, and another portion is transported but retained in the sediment of both lentic and lotic waters, collectively referred to as 'legacy P' [5].

Different from active P which is readily lost by surface runoff and soil erosion, legacy P from NPSs exhibits distinct three-dimensional features throughout its retention-remobilization-hysteresis trajectory along the terrestrial-aquatic continuum (figure 1). Vertically, the accumulated NPS legacy P in soils will percolate into groundwater upon surpassing the leaching threshold. At the water-sediment interface, legacy P also undergoes vertical release governed by adsorption-desorption equilibrium, hydrodynamic conditions, and other environmental factors. Longitudinally, soil legacy P can migrate along hydrological gradients by subsurface flows into rivers, and released legacy P from sediments is transported by advection and diffusion. The P biogeochemical processes (e.g. precipitation, dissolution, mineralization, and uptake) among water, soil/sediment, and biota result in long turnover times and biogeochemical hysteresis. Years to decadal water transit times for subsurface transport of legacy P also delay its impact on receiving waters. Consequently, the interplay between hydrological and biogeochemical mechanisms highlights the hysteresis effect and necessitates the integration of temporal dimension.

The three-dimensional features of NPS legacy P shape its delayed behaviors, obscured pathways, and chronic contributions. This complexity leads to a limited understanding of its sources, distribution, processes, fate, and environmental impacts. In this perspective, we argue that three-dimensional perspectives are necessary to better understand the nature and impacts of legacy P, aiming for sustainable and balanced P management worldwide. This perspective unfolds in three sections. First, we revisit historical findings and highlight emerging knowledge gaps. Second, we call for innovative methodologies beyond the confines of monitoring and modeling approaches. Lastly, we suggest a holistic framework integrating diverse disciplines and approaches to tackling the enduring challenge of NPS legacy P.

To understand legacy P, it is essential to identify its sources and pools. The excessive input of anthropogenic P to soils constitutes the fundamental origin of NPS legacy P. According to data from 2000, 71% of the world's agricultural lands are in P surplus [7]. By 2010, the excess in soil P has reached to an average of 24 kg P ha⁻¹·a⁻¹ in China and 10 kg P ha⁻¹·a⁻¹ in India [8]. The majority of applied P are rapidly held by diverse soil minerals through adsorption, precipitation, and isomorphic substitution, which leads to an extremely low concentration (0.1–10 μ M P) in soil solutions for crop utilization [9]. A half-century analysis of global P cycling suggested that only about 22% of P mineral entered food production [10]. This is attributed to the P biogeochemistry including its high sorption capacities by iron oxyhydroxides or coprecipitation with minerals, resulting in poor mobility and solubility of legacy P [11]. From the soil legacy pool, accrued P is prone to leach into groundwater in which the enrichment of legacy P predominantly occurs through two key processes: the reductive dissolution of P-rich iron oxyhydroxides and the mineralization of organic P [12]. Through the transport of eroded P, sediments in lotic and lentic waters are another important legacy pool to reserve deposited and adsorbed P. Sediments act as both sources and sinks of P, depending on conditions such as oxygen levels, pH, and biological activity [13]. Chronic release of legacy P from these pools occurs along the terrestrial-aquatic continuum in the longitudinal dimension, while the remobilized P may become legacy P again as it travels in river corridors. In contrast to active P, it takes relatively long transit time for remobilized legacy P to reach receiving waters [14]. The coupling of hydrological and biogeochemical processes determined the delayed behaviors and environmental impact of legacy P. As a result, retention, remobilization, and hysteresis of legacy P were intrinsically linked, making it insufficient to consider only one dimension in isolation.

Even previous research provided continuous insights, every finding we made leads us to even more questions for unveiling the unknows of legacy P. First, even though we have observed the significantly different behaviors of legacy P between environmental phases, there is a lack of comparative studies on heterogeneous distribution and loading dynamics of legacy P across different pools. Watershed managers need to identify priority control areas to manage P loss effectively. Therefore, it is crucial to identify the pools that contain the largest legacy loads and to understand how the pool storages vary in space and time. Nevertheless, this aspect is insufficiently addressed. Given this research gap, the apportionment of sources and contributions from diverse legacy pools is also constrained.

While it is well known that global change (e.g. climate change, urbanization, food production, and food trade) impact P cycling, detailed mechanisms of how these aspects affect the shift from steady-stead to non-equilibrium and highly variable responses of legacy P on a broader spatiotemporal scale have not been sufficiently addressed. To illustrate, extreme droughts can lead to soil desiccation cracking that directly alters the soil structure with increased macropores and preferential flow paths [15]. This phenomenon allows water to percolate rapidly, bypassing the soil matrix that normally adsorbs P and thus increasing P leaching into deeper soil layers and groundwater [16, 17]. The impact of urbanization on legacy P was revealed by a reconstruction study integrating lake sediment core, water quality, and historical land use/land cover data. From the early 1990s till today when urbanization greatly accelerated, P loadings to Lake Wilcox drop to its lowest level due to the P retention in urban stormwater ponds and bioswales [18]. It is crucial to determine the extent (longitudinally) and timing (temporally) of the potential loss of the retained P under increased variability of forcings in the future, which remains highly uncertain. Food production and trade are two other aspects closely related to legacy P. Around 82% of the global mineral fertilizer was applied to farmland and pasture [19], indicating that global food production is largely dependent on P fertilizers [20]. However, the persistent P loss and low crop utilization of P fertilizers makes P a scarce resource for food production. A modeling scenario analysis indicated that the current legacy P in French agricultural soils could sustain food production for several decades even in the absence of fertilization supply. This finding highlighted the need to match anthropogenic P supply to crop demand considering the legacy reserves, as well as the optimal use of legacy P for a transition towards sustainable low input agriculture [4]. Food trade can also play a crucial role in managing legacy P by influencing agricultural practices and fertilizer distribution on a global scale. Results estimated that global crop trade could save 0.2 Tg P y^–1 of P fertilizers globally [21], while an optimized multilateral crop trade among China and Central Asia would lead to a significant transition from P surplus to mitigation [22]. Integrating the three-dimensional feature of legacy P into food trade strategies could enhance the efficiency of legacy P use and mitigate legacy P loss, although this has not been thoroughly studied.

Decoupling legacy P from active P is a critical challenge we must overcome for better clarifying what we mean by 'legacy'. A key aspect of the solution involves analyzing the P speciation in soils or sediments to identify which fractions are bioavailable and which are stabilized. Soil phosphorus storage capacity and P saturation ratio (PSR) are two typical site-assessment proxies to predict the legacy P storage and potential of P loss from soils and sediments [23]. At the core of both methods is the concept of P saturation—the point at which the soil or sediment cannot retain any more P. This involves the interdisciplinary fields of soil science and agronomy to balance fertilizer use, crop yields, and legacy P loss.

However, the field-scale methods indeed fall short in estimating the watershed outcomes of legacy P. Watershed-scale methods should consider not only the mass balance of P input, output, and riverine outflows, but also the variability in soil properties, land use patterns, and hydrological pathways across the terrestrial-aquatic continuum. Quantitatively investigating the storage, distribution, and flux of legacy P within watersheds is essential yet challenging, because they are associated with long transit times that tend to evade detection when currently prevailing watershed models are used [24]. A recent modification of the SWAT model enabled explicit sketching of in-stream legacy P hysteresis impact and contribution to watershed yields by adding in-stream P retention and remobilization processes, which reflects the benefit to incorporate insights from different disciplines such as hydrology, geography, and environmental science [25]. Another innovative approach integrated travel time distributions and convolutions to simulate the dissolved P loads leaving legacy pools via subsurface pathways. The developed model traced the accumulation and depletion trajectories of legacy P across the terrestrial-aquatic continuum over a century [26].

Therefore, addressing the unprecedented challenges posed by legacy P requires the integration of multi-scale approaches and multi-disciplinarity. Field-scale soil testing, while offering valuable site-specific data on the concentration and speciation of legacy P, is constrained by its limited scope and spatiotemporal reach. On the other hand, the soil P pools (e.g. labile, active, and stable) conceptualized in widely-used watershed models are not tied to specific P speciation tests (e.g. Mehlich, Bray, and Olsen) that imply the presence of legacy P. The watershed-scale modeling of legacy P also urgently requires knowledges from geochemistry and limnology which can reconstruct historical change of P loading by lake or reservoir sediment cores to facilitate long-term model validation. These critical gaps underscore the necessity for integrating field observations with modeling efforts, moving beyond the current limitations of relying on single methods and disciplines.

Present research has laid the groundwork for a deeper understanding of the three-dimensional trajectory of NPS legacy P. However, the current focuses on single structure or process resulted in a lack of consideration for landscape combinations along the terrestrial-aquatic continuum. An in-depth exploration of the sources, distribution, processes, fate, and environmental impacts of legacy P from a systemic framework and a continuum perspective is urgently needed (figure 2).

The retention-remobilization-hysteresis trajectory of NPS legacy P are manifested not only in the longitudinal and vertical dimensions interactions, but also in the temporal scale due to the lag effects brought by hydrological and biogeochemical coupling processes. The three-dimensional characteristics shape the delayed behaviors, obscured pathways, and chronic contributions of legacy P. This complexity makes tracing, characterizing, and managing legacy P more complicated and challenging compared to active P. Therefore, revealing the multi-dimensional interaction of the retention-remobilization-hysteresis trajectory of legacy P warrants additional research.

Future studies on legacy P should extend beyond its impacts on receiving waters. Considering the estimated low watershed P buffering capacity (2.1 t P km⁻² with a range between 0.03–8.7 t P km⁻²), the legacy pools are unlikely to retain additional P beyond the threshold [27]. These pool storages vary among surplus-depletion-deficit conditions through retention-accumulation-release processes. Therefore, the dynamic behavior and high complexity of legacy pools need in-depth study. Given the intense global change in the future, extreme events, land use transitions, and agricultural intensification may significantly complicate the study of the cycling pattern of legacy P.

We argue that unraveling the complexities of legacy P requires an integration of knowledge and methodologies from various disciplines, such as soil science, hydrology, agronomy, limnology, and environmental science. It also necessitates multi-scale approaches, from the field-scale interactions between legacy P, water, soil, and sediment to the watershed-scale processes governing the retention-remobilization-hysteresis trajectory. Such an integrated framework can enhance our understanding of the intricate interactions among land use, climate change, food production and trade, and P cycling patterns, paving the way for globally sustainable P management over the long term.