Critical Review

Cite This: Environ. Sci. Technol. 2018, 52, 2438−2450 pubs.acs.org/est

Human Perturbation of the Global Cycle: Changes and Consequences † † Zengwei Yuan,*, Songyan Jiang, Hu Sheng, Xin Liu, Hui Hua, Xuewei Liu, and You Zhang State Key Laboratory of Control and Resource Reuse, School of the Environment, Nanjing University, Nanjing 210023, China

*S Supporting Information

ABSTRACT: The phosphorus (P) cycle is an important Earth system process. While natural P mobilization is slow, humans have been altering P cycle by intensifying P releases from lithosphere to ecosystems. Here, we examined magnitudes of which humans have altered the P cycles by integrating the estimates from recent literatures, and furthermore illustrated the consequences. Based on our synthesis, human alterations have tripled the global P mobilization in land-water continuum and increased P accumulation in with 6.9 ± 3.3 Tg-P yr−1. Around 30% of atmospheric P transfer is caused by human activities, which plays a significant role than previously thought. Pathways involving with human alterations include extraction, application, wastes generation, and P losses from cropland. This study highlights the importance of sustainable P supply as a control on future food security because of regional P scarcity, food demand increase and continuously P intensive food production. Besides, accelerated P loads are responsible for enhanced worldwide, resulting in water quality impairment and aquatic losses. Moreover, the P enrichment can definitely stimulate the cycling of carbon and nitrogen, implying the great need for incorporating P in models predicting the response of carbon and nitrogen cycles to global changes.

1. INTRODUCTION calculate P flows or stocks. For example, atmospheric P emissions, deposition and transportation have been estimated Phosphorus (P) is a structural and functional component of all 17−19 lives, and therefore is vital for the existence of humanity.1,2 In using observed data and chemical transport models. A unmanaged and pristine ecosystems, P cycling is considered to stepwise-enhanced model was developed to analyze the spatial “ ” variability in global P release from the lithosphere due to be tight - P transfers between biota and soil within the system 20 are much larger than those gains or losses across the system chemical . Global P transfers within terrestrial boundary.3 Humans have been altering the global P cycle by ecosystems were calculated with a grid-based global geo-

Downloaded via NANJING UNIV on December 30, 2019 at 10:23:58 (UTC). 21 mining phosphate rock, applying commercial fertilizers, and chemical model. The P transportation within the aquatic changing the pathways of P losses. Rising P availability continuumwasestimatedwithgrid-basedandspatially 22,23 fl increases P inputs into terrestrial ecosystems, as well as P distributed models. P ows in agricultural systems and

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. human society have also been estimated from mass balance losses from land to freshwaters and costal zones, consequently, 24−26 turning the “tight cycle” into a much more open cycle.4 In the approaches. However, these estimates focus on parts of short term, P losses can cause eutrophication in waterbodies, the global P cycles and cannot provide a complete perspective. which leads to algal blooms and many other adverse effects, In this study, we conducted a synthesis to provide the such as water quality impairment and fish kills.5,6 Long-lived P magnitude of the extent to which humans have altered the P in soil and sediments can also have long-term environmental cycle based on estimates from some recent studies using “legacy” impacts by chronic P release from these accumulation observed data and spatial-based model. We paid special fl pools.7,8 Given the present trends in and attention to how human activities in uence P cycle by diet preference, the demand for inorganic P fertilizers would including mineral P uses in socio-ecosystem. Further, we increase accordingly in the coming decades, which has also examined the consequences of human perturbation of P cycle − raised concerns about future P security and shortages.9 11 on food security, eutrophication and ecosystem damage, and Changes in global carbon (C) and nitrogen (N) cycle have connection to the cycling of other major elements. This study been well understood,12,13 but human impact on the global P cycle has received relatively little attention (see however Received: August 1, 2017 − Compton et al., 2000; Smil, 2000; Bennett et al., 2001).14 16 Revised: January 25, 2018 However, more recently, there has been an expansion in Accepted: February 5, 2018 studies, using global observed data and spatial-based models to Published: February 5, 2018

© 2018 American Chemical Society 2438 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Figure 1. A schematic diagram of (a) background state of global P cycle and (b) human altered global P cycle. The blue arrow indicates phosphorus flows linking land and . Red arrows indicate P flows linking human and natural ecosystems. The bold black figures in Figure 1a show the magnitude of P stocks. It should be noted that the P transfers by riverine runoff are net result of natural , soil loss, wastes and retention. The unit of P flux is Tg P yr−1 and the unit of P stock is Tg P. The definitions of the symbols in the figure can be found in SI Table S3. closed with a discussion on several key issues that remain to be Engine with a combinations of the following terms in title: solving and intervention strategies for reducing human impacts (global OR world OR earth OR globe OR planet) and on global P cycle. (phosphorus OR nutrient OR P OR OR phosphate) and (terrestrial OR land OR soil OR water OR river OR lake 2. THE GLOBAL P CYCLE AND HUMAN ALTERATION OR aquatic OR ocean OR marine OR sea OR atmosphere OR biota OR plant OR animal OR sediment OR reserve OR stock The cycling of P is unique in that it consists of a massive and fl wholly unavailable pool of apatite minerals in the lithosphere; a OR pool OR deposit OR mineral) and ( ow OR cycle OR relatively small atmospheric transfer; a relatively simple cycling OR budget OR release OR transport OR model OR transfers of P that cycles among terrestrial ecosystems, and retention OR deposition OR weathering OR emission). between land and ocean through riverine runoff. The most Besides, we also conducted a search in google engine to yield important anthropogenic change to the global P cycle is a potential reports using general terms including (UN OR USGS massive transfer from the vast and unavailable reserve pool to OR IFA OR FAO) and (phosphorus) and (mineral OR reserve biologically available forms on land. To get a comprehensive OR sediment OR stock). At last, a total of 63 publications that estimate of the human alteration on P cycle, we analyzed the contain quantitive information on global P pools and flows peer-reviewed primary articles searched using Google Scholar were compiled in Table S2 in Supporting Information (SI).

2439 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Figure 2. Global P production and consumption. (a) Historical trends of production of phosphate rocks and consumption of P. (b) Share of P production and consumption in 2014. Data are from online database provided by Kelly and Matos (2014) and FAO (2016).38,39

fi ± −1 15,23 −1 From these publications, we identi ed eight P pools (sedi- 6.5 1.5 Tg P yr (F5). The higher value (8.0 Tg P yr ) ments, land, mineral reserves, terrestrial biota, oceanic biota, was estimated at by multiplying an extrapolation of global freshwater biota, freshwater, and seawater), and complied 23 P erosion rate and global averaged soil P content.15 The lower flows between these pools (SI Table S3). In this study, we bound value of 5.0 Tg P yr−1 was provided by a more recent followed the expression format of results from Wang et al. estimate based on a biogeochemical model (named IMAGE- (2015), which provided a most likely estimate and a 90% GNM) that considered spatial variability of climate, hydrology confidence interval.19 There are substantial discrepancies in the and human activities.23 P transfer to the ocean via riverine ff ff ff ± −1 15,23 estimates from di erent existing studies because of the di erent runo was estimated at 4.5 1.5 Tg P yr (F6). methods and assumptions used, as well as differences in Subtracting this amount from P inputs to freshwater, the P considering local nuances caused by human activities, like farm retention rate in freshwater is estimated at 2.0 ± 1.5 Tg P yr−1 27 operations and conservation strategies. Here, we conducted a (F7). Once entering ocean, most of P exists in sediments in uniform synthesis based on a semiquantitative method to estuaries and coastal areas with a rate of 2.0−3.0 Tg P yr−1 fi ff ± obtain the most likely estimate and 90% con dence interval (F10). In contrast to P transport via riverine runo , only 0.5 from existing results. In general, we first assessed the reliability 0.5 Tg P yr−1 is transported to ocean through atmospheric ± −1 of the results from existing studies using on an indicator-based transfer (F3), as most part of 1.7 0.3 Tg P yr (F2) tends to method. Then, based on the reliability, we decided the most redeposit near its origin due to the short residence time of dust likely estimate and the 90% confidence interval as the estimates in the atmosphere.18,19,31,32 of P pools and flows in this study. Finally, we assessed the Transfers in Organic Cycles. The land- and water-based reliability of the estimates based on the reliability of estimates organic cycles transfers P between environment and living we used. The detailed method can be found in SI Text S1−S3 organisms, and through living organisms via food web. The and Table S1, and the results can be found in SI Table S4−S5. land-based organic P cycle consists of P moving from to The Figure 1 shows the final estimates of changes of P cycle, plants and animals, and then returning back to soil, where it which are also described in following sections. becomes available again for plant uptake through mineraliza- 2.1. Background State of P Cycling. Initial Source. The tion. P stock in global soil pool has been estimated to between largest P pool on the Earth is the apatite minerals in the 0.23 × 105 and 2.0 × 105 Tg P. The nine factor between the lithosphere, which is dominated by the oceanic and freshwater lower and higher bound estimates was sourced from the sediments. However, almost all the apatite minerals in the pool, methods used in these study, while the higher values (0.96 × with an amount of 0.8 × 109 to 4.0 × 109 Tg P, lies beyond the 105 to 2.0 × 105 Tg P) were obtained based on global averaged 28,29 reach of plants (R1, Figure 1a). The initial biota available P approach neglecting spatial heterogeneity of soil P con- to the Earth system is from the P releasing from apatite tent,29,33,34 and the lower values (0.23 × 105 to 0.59 × 105 minerals via chemical weathering in soil formation process, Tg P) considered spatial distribution of soil order and different which varies greatly with rates between 50 and 720 g P m−2 P content in each order.21,35 P stock in terrestrial biota was −1 30 ± × 2 yr across the world. The recently estimate of global estimated at (4.7 0.8) 10 Tg P (R4). In the early year, this chemical weathering rate was 1.5 ± 0.4 Tg P yr−1 by spatially sock was estimated based on a combination of global C stock in explicit model simulation (SI Text S3.2).20,21 In contrast, standing (4.5 × 105 to 8.3 × 105 Tg C) and molar C/P physical weathering produces fine materials that are typically ratio of 500−882:1.28,29,33,34 However, Smil (2000) pointed out unavailable to biota, but prepares raw material for chemical that these estimates (varying from 1.8 × 103 to 3.0 × 103 Tg P weathering, which was estimated at a rate of 10.0−15.0 Tg P yr−1) were too high molar C/P ratio used in these studies yr−1 roughly based on continental denudation and crust P represents C/P ratio in leaves, which has significant higher P 16 15 content (F4). content than that in the whole above-ground biomass. He Transport from Land to Ocean. Transfer of eroded soil by provided an estimate of 5.5 × 102 Tg P using a much higher riverine runoff delivers P from land to ocean. This process starts molar C:P ratios of 3745−5244:1. This estimate has been from the soil-derived P inputs to freshwater by erosion and supported by a more recent estimate of 3.9 × 102 Tg P, using a runoff. In nature state, P input to freshwater were estimated at 0.5° by 0.5° global biome database and detailed P/C mass ratio

2440 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Figure 3. Phosphorus use in crop farming. (a) Historical trends in global P input and P use efficiency (PUE) in global crop farming. Data are from studies by Bouwman et al. (2013) and Chen and Graedel, (2016).25,26 (b) Historical trend of P inputs intensity, PUE and dependence on chemical P fertilizer (DCP) of China. Data are from study by Liu et al. (2016).47 (c) Relation of economic development level and PUE. Data are from references listed in SI Table S7. (d) Relation of economic development level and DCP. Data are from references listed in SI Table S7. in leaf, wood and root of 13 plant types.21 As the P flow for fertilizers from the 1940s to the 1980s. Thereafter, between soil and terrestrial biota was estimated at (0.8 ± 0.2) × extraction declined slightly for several years due to the collapse 2 −1 15,21 10 Tg P yr (F8, Figure 1a), the residence time of P in of the Soviet Union and the decreased fertilizer demand in land-based organic cycle is estimated at 5−6 years on average developed countries in Western Europe and North America.39 by dividing the estimated P stock in terrestrial biota and P flux. From 2000 onward, the rapidly growing fertilizer demand in Estimates of P transfers in organic P cycle within the ocean developing areas has seen another surge of phosphate rock are better constrained. P stocks in seawater and oceanic biota extraction, with an annual growth rate of 4.2%. are estimated at (1.0 ± 0.2) × 105 Tg P and (1.0 ± 0.3) × 102 Phosphate rock is first refined into phosphoric acid, either by 40 Tg P (R8 and R5), and P transfer rate between the two pools wet process or the thermal process. Almost all the wet ± × 3 −1 15,28,29,33,36 was (1.0 0.2) 10 Tg P yr (F9). Similarly, the process-based phosphoric acid are used to produce fertilizer. residence time of P in oceanic-based organic P cycle is ∼0.1 Global fertilizer production is skewed in a handful of countries, year, which is more efficient than land-based organic cycle. with China and the United States at the top, accounting for Estimates of organic P transfers in freshwater are not very more than one-half of the world production (Figure 2b).39 common. The only estimate of P stock in freshwater biomass However, most thermal process-based phosphoric acid is for fi and P uptake by freshwater biota was 0.34 Tg P (R6) and 10 Tg ne chemicals, which are further processed into food/feed −1 33 fl Pyr (F10) by Pierrou (1976). The residence time of additives, detergents, pesticides, surfactants lubricants, ame freshwater based organic P cycle is of the order of ∼0.3 year. retardants, and metal treating products (Villalba et al., 2008). 2.2. Anthropogenic P Cycles. Human alterations of the P Despite the widespread use in many industries, P-containing cycle involve with a number of pathways, including phosphate fine chemicals only presented around 20% of extraction.26 rock extraction, crops cultivation with fertilizer application, Crops Cultivation with P Fertilizer Application. The livestock production with manure generation, and food traditional crop farming relied on P in soil and local from consumption with wastes production. recycled organic wastes, which consisted the main part of Phosphate Rock Extraction. The extraction of phosphate agricultural pattern for thousands of years.41,42 However, an rock from mineral pools is the primary source of human exponential increase of inorganic fertilizer application was alterations on the global P cycle. By now, a total of 1.1 × 102 Tg witnessed in crop farming from 2.0 Tg P yr−1 in the 1940s to P in phosphate rocks has been extracted from mineral reserves 13.8 Tg P yr−1 in the 1980s driven by the consumption since the 1840s when production expansions in developed countries (Figure 2a).38,39 Since the technology was first applied to produce ordinary super- 2000s, the expansion of fertilizer consumption in developing phosphate in the UK (Figure 2a).37,38 The extraction of countries contributed to another rapid increase of inorganic P phosphate rocks slowly increased over 2 Tg P yr−1 by the year fertilizer application with an annual growth rate of 2.5%, and ± −1 1940, and experienced a 10-fold increase because of P mining now 17.1 2.3 Tg P yr was applied to agricultural land (F13,

2441 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Figure 4. Phosphorus inflows in domestic sector. (a) Categorical classification of P inflows in domestic sector. Data are from Chen and Graedel, (2016).26 (b) P outflows from domestic sector in several countries. CN: China. JP: Japan. KP: South Korea. FR: France. DE: Germany. FI: Finland. SE: Sweden. AT: Austria. CH: Switzerland. NL: The Netherlands. Data are from references listed in SI Table S7.

Figure 1b). Meanwhile, P uptake in harvested crops have availability of inorganic P fertilizers.11 Averaged globally, only ± −1 26,43−46 ± −1 − reached to 12.3 0.3 Tg P yr (F14). However, the 10.2 4.8 Tg P yr in manure (F15), about 50 60% of the growth rate of P uptake in harvested crops (1.0% annually) was total amount, was returned to cropland.26,45 Note that the much lower than that of P fertilizer application, indicating a manure P rate has great spatial differentiation because remarkable decrease of P use efficiency (PUE) in crop farming of various local factors such as breeding patterns, cultivation (Figure 3a). This is highly related to the increasing dependence technologies, agricultural management practices and environ- on chemical P fertilizers (DCP), which accounted for mental awareness.70 In West European and North America, approximately 60% of the total P applied in crop farming in 73−82% of P in livestock are recycled, whereas it is 30−50% in 2013 (Figure 3a). The decrease of PUE is more significant in other regions, like Africa, Asia, and South America.45,70 some developing countries, like China, who consumed 28% of However, the fate of P contained in manure that have not ± −1 global P fertilizer consumption (Figure 2b). In China, P been recycled (8.3 4.3 Tg P yr ,F22), has not been well- application rate increased from 5.4 to 77.5 kg ha−1 between understood. This huge amount is an important “missing piece” 1949 and 2012, accompanied by a continuous increase of the related to animal breeding in global P cycle, which could act as DCP and a continuous decrease of PUE (Figure 3b).47 an important source of nonpoint P pollution.71 However, decreasing PUE has begun to reverse in some Human Consumption and Wastes Generation. P con- developed countries. At the global scale, countries with higher sumption by human directly has tripled since the 1960s (Figure −1 gross domestic product (GDP) per capita tend to have higher 4a), and now reached to 9.9 Tg P yr (F16 and F17), with about PUE as well as less DCP, as PUE is positively correlated with 60% of the increase coming from the use of mineral P economic development level (R2 = 0.53, P < 0.001, Figure 3c), products.26 Processed food (6.2 Tg P yr−1) was the major while DCP is negatively correlated with GDP per capita (R2 = source, but the detailed P inventory of chemical products is still − 0.55, P < 0.001, Figure 3d).47 66 The only exception is Uganda, unclear because of the trace amounts of P in them and the an African country with low DCP and high GDP per capita, widespread use. Globally, the annual average of P in per capita which represents many African countries where farmers cannot food consumption was 0.86 kg P capita−1, of which 0.40 kg P afford inorganic P fertilizers. capita−1 ended up as municipal wastes averaged globally.26 The Livestock Production and Manure Generation. Globally, P amount of P ending up in waters in some developed countries inputs in feeds increased 2-fold in the past five decades, while P was relatively higher compared to global average, such as 0.63 in livestock products (e.g., meat, milk, egg) tripled from 0.5 to kg P capita−1 yr−1 in France.58 The annual average P intake 1.6 Tg P yr−1, illustrating a slight improvement in PUE of (0.48 kg P capita−1) was much higher than recommended livestock farming,26 with 8% of P being converted to livestock dietary allowance (RDA) of 0.26 kg P capita−1 for adult aged products now. The main carrier of the rest P is animal excreta, over 18, but was equivalent to that for teenagers aged 9−18, with an amount of 15.0−24.0 Tg P yr−1, equaling to inorganic P who have higher P demands to build their bodies.72 fertilizer application in crop farming.25,26,67 However, some As 98% of the daily ingested P is excreted, 7.2 billion people regions with high livestock densities are facing an excess of across the world would generated 3.3 Tg P yr−1 of excreta in manure since the livestock farming tends to shift from total.43,73 Currently, only around 20% of human wastes are 67 ± −1 26,33,43 distributed smallholders to intensive operations. For example, recycled, with a magnitude of 1.3 0.2 Tg P yr (F19). in China, the world’s leading pig producer, the intensive In premodern East Asia, it was a common practice for at least breeding systems contributed to 64% of the national total pig 5000 years to collect and recycle human excrement.74 Especially production in 2010, while it was 5% in the early 1980s.68 in China, more than 90% of human excrement was recycled to Globally, about 70% of poultry and 58% of pig products came grow crops until the 1980s.75 But now, only 60% of P in human from intensive systems.69 With livestock being clustered in excretment is returned to cropland in rural areas, and it is only specific areas, over 50% of the global manure P was produced in 10% in city areas.76 Possible reasons include but not limit to the 13% of crop farming areas, mainly in China, India, the widerspead use of flush toilets and municipal seweage systems. Northeastern America, EU27, and Brazil.67 In many cases, it is Spatially, the average recycling rate of P-containing human economically impractical to recycle livestock manure beyond a wastes in East Asian countries, about 24% (20−27%), is about certain radius because of high transport costs and the twice of that in European countries, estimated at 12% (5−20%)

2442 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Table 1. Estimates of Longevity of the World’S Phosphate Reserves

longevity depletion reference (years) (year)a assumption Smil (2000)15 80 2080 based on “current extraction rates” Fixen (2009)95 93 2102 based on a constant extraction rate in 2007−2008 Smit (2009)97 70−100 2079−2109 assumed a range of 0.7−2% increase in production until 2050 and 0% thereafter Vaccari (2009)98 90 2099 based on “current extraction rates” Van Vuuren (2010)96 90 2100 based on scenario analysis considering population size, economy condition and agriculture expansion Van Kauwenbergh 300−400 2310−2410 based on “current extraction rates” (2010)102 Koppelaar (2013)101 300−400 2300−3400 based on scenario analysis considering population size, food demand, waste recycle and fertilizer price aEstimated from the lifetimes and date of publications.

(Figure 4b).47,48,50,51,56,58,61,62,65 As for the P contained in contribute to the buildup of P pools in agricultural soils. By domestic wastes, a small amount, was discharged to freshwater aggregating all the estimates of P fluxes through croplands, we ± −1 ± ± −1 (1.1 0.5 Tg P yr ,F18) and recycled to agriculture (1.3 0.2 estimated P accumulation in cropland to be 6.9 3.3 Tg P yr −1 ± Tg P yr ,F19). The remaining was estimated to be 7.6 0.2 (SI Table S6,F23), which consists with a previous estimate of 9 −1 fi −1 24 Tg P yr (F21), by balancing inputs (F16+F17) and quanti ed Tg P yr . Globally, more than 70% of the global cropland “ ” 45 outputs (F18+F19). This important missing piece in global P area faces P surplus issues. In China, P stock in arable land cycle could ended in landfill or abandoned without disposal, underwent a long-term depletion up to the 1960s and then and may have broader environmental influence once P in reversed into net P accumulation thereafter, with an annual abandoned wastes leaks. accumulation rate of 41 kg P ha−1 now.47 In certain agricultural 2.3. Alteration of Global P Cycling. Accelerated P watersheds, P surplus situation is more severe (74 kg P ha−1 Transfers in Land-Water Continuum. The anthropogenic yr−1).87 It is estimated that global P accumulation in cropland changes in the P cycle drive regional and global changes in the has a magnitude of 420 Tg during 1965−2007, and was aquatic P transfers. P losses from cropland through erosion and anticipated to increase by 600 Tg in the next 50 years.88 runoff, also known as nonpoint sources, have been recognized However, a lack of our knowledge of the fate of cropland P as the dominant contributor.71,77,78 Globally, around 56% of accumulation makes such a large P stock “a missing piece” in P aquatic P inputs are induced by surface runoff from cropland, cycle related to crop farming. One of the potential fate of this an increase of 21% since the early 1900s.23 Global P losses from “missing piece” is that they are reactivated and absorbed by cropland to freshwater has been estimated to be between 5.0 crops as nutrient, which has been proved at field, farm and − and 26.4 Tg P yr−1,15,23,79,80 with particularly large uncertainties country scale.89 91 For example, a field-scale study found that P − because of variable soil erosion rates and soil P content.81 83 accumulation in a 28-year build-up soil was sufficient for crop Global P losses via erosion and runoff was first estimated at cultivation for 15 years without compromising production.89 15.0 Tg P yr−1 by multiplying global averaged soil erosion rate Increased Atmospheric P Transfers. Atmospheric P trans- (20 t ha−1 yr−1) and soil P concentration (0.05%).15 Quinton et fers were previously thought to be less affected by human al. (2010) provided a large range of of 14.9−26.4 Tg P yr−1 activities, which were estimated to emit ∼0.05 Tg P yr−1 based using a similar method but using soil erosion rate from a on global black carbon emission and P/black carbon ratio.17,92 spatially explicit evaluation.79 Beusen et al. (2016) developed a However, the total atmospheric P emission (1.39 Tg P yr−1) spatially explicit model that considers the spatial heterogeneity based on this value is less than 50% of global atmospheric P of the soil erosion rates and soil P concentrations based on a deposition (3.7−4.6 Tg P yr−1) extrapolated from in situ series of global spatial database, including distribution of observations.18,31 A recent global fuel-combustion inventory precipitation, temperature, surface runoff, land cover, fertilizer estimated that P emission from combustion sources was 1.8 Tg application, and soil P content.23 This study provided a much Pyr−1,19 which helps to explain the gaps between previous smaller estimate of 5.0 Tg P yr−1. More recently, Lwin et al. estimates of global atmospheric budget. In this study, global (2017) provided a similar estimate of 5.7−6.1 Tg P yr−1 using a atmospheric P emission was estimated to be 3.9 ± 0.7 Tg P numerical model.80 Based on our synthesis, the estimate of P yr−1, 3.0 ± 0.3 and 0.9 ± 0.5 Tg P yr−1 of which was estimated ± −1 losses from cropland to freshwater is 10.4 5.7 Tg P yr (F20, to deposit over land and ocean, respectively (F1,F2, and F3, Figure 1b). By subtracting it from the total P inputs to Figure 1b).18,19,31 Some studies pointed out that human freshwater (18.0 ± 9.0 Tg P yr−1),15,23 P losses from natural activities, including P fertilizer addition, reclamation, livestock ± −1 erosion is estimated to 7.6 3.3 Tg P yr (F5, Figure 1b). , and , may also increase atmospheric P With the increase of P inputs to freshwater, P entering ocean mobilization by increasing soil P content and wind erosion.93,94 via riverine runoff also had an increase, which was estimated to Zhang et al. (2017) estimated that global averaged P content in ± −1 15,23,33,84−86 46 be 12.6 5.6 Tg P yr (F6, Figure 1b). By cropland soil increased by 15% in the last hundred years, balancing this value with total P inputs to freshwater, the P which could cause a ∼15% increase in P transfers to the retention in freshwater is estimated to be 5.4 ± 3.2 Tg P yr−1 atmosphere from cropland via wind erosion if other factors are (F7, Figure 1b). However, the fate of P retained in freshwater not considered. Multiplying 15% with the proportion of has not been well understood, making it a “missing piece” in cropland on total land area (11%),39 we estimated that the global P cycle. increase in cropland soil P content could only increase 1.7% Enhanced Soil P Accumulation. As P compounds are lowly global atmospheric P transfers via wind erosion. Thus, the mobile and tend to confine, the large agricultural P inputs underestimate of wind erosion from improved cropland soil P

2443 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Figure 5. Top exporters (a) and importers (b) countries of mineral P and their cumulative share of the global total amount trade. Data are from the online database provided by IFA (2016).104 Morocco: MA; China: CH; Russian: RU; The United States: US; Jordan: JO; Saudi Arabia: SA; Peru: PE IN: India; BR: Brazil; Canada: CA; Indonesia: ID; Australia: AU; Argentina: AR; Pakistan: PK.

Figure 6. Loads of TP (a) and DP (b) to waterbodies as a function of human in this watershed. TP and DIP loads are from Mayorga et al. (2010),86 and the selected watersheds accounted for the top 55% of global total loads. Population data are from CIESIN (2015) and reprocessed in ArgGIS.111 content (0.017 Tg P yr−1, by multiplying 1.7% with current Africa with Morocco alone holding 73% of the global share.103 − estimate of wind erosion of ∼1TgPyr 1 from ref 19) may not International trade has been playing an important role in have a significant effect on global atmospheric transfers. redressing regional P scarcity. One third of the global phosphate rock production was spread to the rest of the 3. CONSEQUENCES OF HUMAN ALTERATION world from these top seven suppliers, accounting for 82% of the total traded P, and 58% of P trade went to top seven importers 3.1. Phosphorus Scarcity and Food Security. The rapid (Figure 5a, 5b).104 However, with the depletion of phosphate increase in phosphate rock extraction makes global P scarcity reserves in suppliers, like China and America, P supply would worth paying attention, as it takes millions of years to reform fi become more spatially concentrated with Morocco providing phosphate rocks, making P a nite resource in the civilization 105 time scale. Global P scarcity is often measured from the provide 80% of global phosphate rock by 2100. longevity of phosphate reserves by dividing the world’s total As there is no substitute for P in food production, P scarcity phosphate reserves by the annual P consumption and has been is directly linked to future food security determined by several 15,95−98 factors. First, as the world’s population is projected to increase estimated by a number of studies (Table 1). After the − 106 fi United States Geological Survey (USGS) updated the estimate by 32% during 2015 2050, there is de nitely an increase in of global phosphate reserves from 16 000 Tg to 65 000 Tg in food demand. Tilman et al. (2011) anticipated a growth rate of − 9 2011,99,100 the longevity of phosphate reserves was recently 100 110%. Second, global diets are changing toward more − estimated 300−400 years,101,102 which indicated that the world meat and dairy products. Animal derived food requires 7 10 as a whole would not face global P scarcity for the next couple times more P in the form of phosphate rock than that crop 87 centuries. requires from a life-cycle perspective. Therefore, the changing Even though the global phosphate reserves are not about to diets imply an growing P demand in meeting per capita food be depleted in a short-term, the regional P scarcity would demand, which has been observed to increase by 38% (from 1.9 to 2.6 kg P capita−1 yr−1) over the past five decades, and would probably become more severe and cause pressure on future − − food security. Regional P scarcity presents P demand in many rise up to 3.7 kg P capita 1 yr 1 by 2050.107 Third, food regions, especially some populous countries like India and production system is becoming more P intensive, as modern Brazil, depends on a small number of P-rich rock producers, as agriculture is more dependent on phosphate rock and is more than 85% of global phosphate reserves are located in five becoming less P efficient, especially in regions with excess countries, namely Morocco, China, Algeria, Syria, and South fertilizing, like China.108 Based on scenario analysis, Van

2444 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Vuuren et al. (2010) estimated that the global P demand for Eutrophication has substantial impacts on ecosystem food production would continuously increase to 37−48 Tg P function in fresh waters. A most instant effect of eutrophication − yr 1 around the year 2100, nearly twice of that in 2015 (20.0 Tg is the rapid growth of phytoplankton species. In extreme cases, 96 P). Using the Millennium Ecosystem Assessment cereal yield it can turn into blooms of harmful algae (HABs), killing fishes storyline and current P application rates, Peñuelas et al. (2013) and producing harmful metabolites that threat drinking water estimated that over 50% of global mineral P pool will be security. HABs were responsible for the drinking water crisis of fi 109 depleted by midlate twenty- rst century. Accordingly, the nearby city in Lake Tai in 2008 and in 2014 in Lake Erie.124,125 threat of P on food security still warrants special attention. Eutrophication is also associated with the hypoxia in the 3.2. Phosphorus Loss and Eutrophication. Human- hypolimnia of waters. A studies of 11 eutrophic lakes suggests induced P losses contribute to eutrophication in two major that this is associated with the decomposition of settled organic waysimmediate P exports from human activities does so material at the sediment surface and oxidation of substance directly, and chronic release from enhanced P pools can be a diffusing from the sediment.126 Using the records of hypoxia in further source of eutrophication. Given human-induced ∼ acceleration of P losses, it is no surprise that P loads and 365 lakes worldwide, Jenny et al. (2016) found that 20% of concentration in surface waters have increased over time. An the total lakes have shifted to hypoxic condition since the estimate of total P (TP) loads to the world’s 100 largest lakes middle of the 19th century, and the number was well correlated 2 127 indicates that the average loading was 7% higher during 2005− to total P release (R =0.99, P < 0.001). In the long term, 2010 than that during 1900−1995 and this ratio reached up to eutrophication can cause the loss of biodiversity, both in the 79% in South America.110 Moreover, both TP load and composition and richness of plankton and macroinvertebrate. dissolved inorganic P (DIP) loads to waterbodies are correlated Vonlanthen et al. (2012) reported that the species loss of with population densities in the watershed (R2 = 0.50 and 0.88, whitefish in 17 European can be explained by the magnitude of P < 0.001, Figure 6a and b).86,111 However, the correction level eutrophication measured by P concentration (R2 = 0.50, P < of DIP is larger than TP, illustrating that human activities have 0.001).128 a great impact on DIP load than TP load. The enhanced P 3.3. Phosphorus Coupling with Other Biogeochemical loads from watersheds led to increasing P concentration in Cycles. The dynamics of P cycling is associated with the water bodies. Continuous monitoring data in three large biogeochemical cycles of carbon (C) and N,129 through its − shallow lakes in the Yangtze River basin, China has shown that determining role on primary production and N fixation.130 132 TP concentration, China have increased by 3- to 7-fold since 112 The coupling characteristics have also been found by a number the 1950s. The chronic release of P from legacy pools, also 133 134 ff of experimental observations in forests, wetlands and named legacy P e ect, is another contributor to eutrophica- aquatic ecosystems.135,136 To examine the universality of the tion.113 Sharpley et al. (2013) reviewed some case studies about response, Peng et al. (2017) conducted a meta-analysis of the time scales of P recovery after remediation in terrestrial, 7 available experimental observations, and found a systematical river, and standing water environments. In the extreme case of increase of terrestrial C and N pools at 10−23% followed by P Loch Leven Lake in the UK, they found that it took this shallow 137 lake 30 years to recover from P pollution after 60% of the and N&P additions. Using a similar method, Li et al. (2016) external P inputs were prevented during the late 1970s. The also found that P addition increased aboveground and resurgence of eutrophication in Lake Erie have led to several belowground biomass production by 34% and 13%, implying fl studies that examined the role of legacy P effect on that C and N uxes in terrestrial ecosystems are also stimulated 138 eutrophication.8,114 These studies showed that the legacy P by enrichment of P. Global model study found that P effect led to a step-change increase in P of surface waters in deposition contributed to 2% of change in C storage in global Lake Erie basin after the nutrient mitigation target had been forests during 1997−2013.139 achieved for decades. Even if the strictest conservation practice The faster increases of P inputs than N inputs in agronomic like ceasing fertilizer application is conducted, it may take years fertilizer application resulted in the decrease in N:P ratio of the to see decreases in P concentration in Lake Erie. In some total inputs in agriculture areas. In China, a decrease of ∼30% regions, such as the Yangtze River basin, as P accumulation in in molar N:P ratio of input flows, from 20.3:1 to 14.5:1 during 115 landscape pools still experienced dramatic increase, the 1980−2005, has been found in the croplands.140 This alteration ff implications of legacy P e ect is particularly important for the also decreases the C:P and N:P ratios with lower C:P and N:P local eutrophication management strategies. ratios being observed in soils and freshwaters in agricultural As the most immediate and well-understood environmental 141−143 fi areas. While C and N are highly mobile through gaseous consequence of P enrichment, signi cant eutrophication forms, P compounds are inactive and much less mobile. These happened in the Great Lakes of North America in the 1950s, differences confer on human-induced P inputs confining near and now has been a widespread issue in many lakes, estuaries its origin and C and N inputs globally widespread, causing an and coastal areas around the world.116,117 While in most overall increase of C:P and N:P ratios in the vast terrestrial and estuaries and coastal sea, N is the key element controlling the aquatic ecosystem,109,144 causing long-term P limitation on C net primary production (NPP), P is the key limiting factor for 30 eutrophication of many fresh waters.5,118,119 In addition to and N cycles. This has important implications to models fl incorporate P in simulating the responses of biogeochemical nutrient enrichment, the occurrence and the in uence of 145 eutrophication is also influenced by many site-specific factors, cycles of C, N and P to global changes. For example, taking like hydrodynamics, and freshwater type and species P cycling into consideration, land C uptake during 1860−2100 − 129 groups.120 122 A study found that it would underestimate the was estimated to 16% lower based on a C−N−P model, and global eutrophication risk of agricultural P emissions by up to 1 the P demand (924−2110 Tg P) to meet the projected order of magnitude if the above spatially related factors were increases in NPP was significant higher than the result (−89 to not considered.123 262 Tg P) simulated without P cycling.146

2445 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

4. CONCLUSION AND OUTLOOK ecosystems in more concentrated does, causing so-called international pollution transfer, but it can also help to replenish The compelling evidence shows that human beings are altering ffi the P cycle with tripling the P mobilization in global land-water soil fertility in regions that do not have su cient P to sustain continuum and substantially increasing soil P accumulation, and the population. As the international trade are growing in the perturbed P cycle is affecting the sustainability of human importance, assessing the consequences of international trade through ecological and environmental risks in the short term, on global P cycle becomes more urgent. Another is identifying 11 and illustrating the key processes controlling interactions of P and food security challenges in the long term. In the with the biogeochemical cycles of C and N. As the interactions foreseeable future, both the extent and consequences of human of P with C and N are involved with many aspects, alteration of the P cycle is set to increase, considering the need understanding the key process can help for incorporating P of expanding agriculture to fulfill food demand for population into the coupled C−N−P models in a terse and graceful way. growth and diet improvement. Thus, interventions are Explore the Methods to Improve Sustainable P Use. necessary for minimizing the negative consequences while Although there are many techniques to improve PUE in crop optimizing the use of P. Potential options include (1) and animal farming, including conservation tillage, precision increasing P efficiency in food production systems by fertilization, gypsum application, rotation, phase feeding, and improving agricultural techniques; (2) reducing overuse of P genomic selection technology,149,150 the problem is how to fertilizer through policy measures; (3) promoting practical promote them effectively in many developing areas, like China, recovering technologies to recycle P from manure, sludge and India, and many countries in Africa, where PUE in agriculture is kitchen residues; and (4) guiding diet preference with still low. Exploring low-cost P recycling techniques is also a reasonable animal derived food consumption. solution, as current P recycling techniques like large-scale The results reviewed in this study benefit from and help cocompositing, sewage sludge and incineration ash recovery advance the estimation of the magnitude of the global P cycle. techniques are economically costly, which impedes the However, there are still several key issues that remain to be widespread application of P recycling. solved. Improve the Fundamental Science of P Cycling and Methodology for Quantifying them. First, there are still a ■ ASSOCIATED CONTENT “ ” number of missing piece in global P cycle, referring to the *S Supporting Information fi fl fi unquanti ed P pools and ows, and those have been quanti ed, The Supporting Information is available free of charge on the but the fates have not been well understood, like P in livestock ACS Publications website at DOI: 10.1021/acs.est.7b03910. manure and human wastes that have not been recycled, and P ± −1 Detailed description of method, critical examination and retained in freshwater and cropland soil (28.3 11.0 Tg P yr − in total, Section 2.2 and 2.3). To illustrate the role of those data source for the synthesis. Text S1 S3 method and − “missing piece”, it is necessary to explore the knowledges about critical examination for the synthesis. Table S1 S3 the transformation and migration mechanisms of them. For reliability assessment indicators, literature list and − example, to understand the “missing piece” of P accumulation symbols used in Figure 1; Table S4 S5 compilations in soil, long-term field experiments should be conducted to of existing estimates of P pools and flows; Table S6 illustrate the pathways of the P after being applied to cropland. phosphorus budget of global cropland; Table S7 Besides, estimation of P cycle at such a large scales at the globe phosphorus balance in different countries (PDF) requires an extreme volume of data from exploring and integrating multidisciplinary approaches. For example, to ■ AUTHOR INFORMATION reduce the uncertainty in modeling global P losses from cropland, observation data about soil erosion rate using Corresponding Author agronomic approach is needed to test the factors influencing *Phone/fax: +86 025 89680532; e-mail: [email protected]. erosion rates; high-resolution global land-use map and digital ORCID elevation model from remote sensing is needed to provide data Zengwei Yuan: 0000-0002-6533-4170 related natural factors; information about local agricultural fl Author Contributions management practice is also necessary to simulate in uence † from human activities, which requires knowledge from Z.Y. and S.J. contributed equally. empiricists who focus on human activities. At last, teamwork Notes and networking at global scale are also necessary to improve the The authors declare no competing financial interest. accessibility of those data by constructing global data sets of P concentrations, which can reduce the large uncertainties in the existing estimates from uneven distribution of P in sediment, ■ ACKNOWLEDGMENTS soil, waterbodies, atmosphere, biota, etc. For instance, Global This work is financially supported by the National Key River Chemistry database (GLORICH) provides data for Research and Development Program of China testing global aquatic P dynamics, and offers a good example of (#2016YFC0502801), the National Natural Science Founda- the role of the data sets in simulating P retention in river by tion of China (#41401652), and the Jiangsu Science dam construction.147,148 Foundation (#BK20140605). The China Scholarship Council Illustrate the Consequence of the P Cycle Changes in (CSC) provided the primary author (Jiang) with a scholarship More Aspects. The first one is assessing the influences of to support a joint Ph.D. studentship at Centre for Ecology & international trade on P cycles. In contrast to transport under Hydrology in the United Kingdom. We thank Helen P. Jarvie the force of nature, the effects of international trade are more from Centre for Ecology & Hydrology for her suggestions and direct and complicated, as it typically results in injection of P to the anonymous reviewers for their helpful comments.

2446 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review ■ REFERENCES lithology, temperature and soil properties. Chem. Geol. 2014, 363, 145−163. (1) Westheimer, F. H. Why nature chose . Science 1987, (21) Wang, Y. P.; Law, R. M.; Pak, B. A global model of carbon, 235 (4793), 1173−1178. nitrogen and phosphorus cycles for the terrestrial . (2) Elser, J. Phosphorus: a limiting nutrient for humanity? Curr. Opin. Biogeosciences 2010, 7 (7), 2261−2282. Biotechnol. 2012, 23 (6), 833−838. (22) Harrison, J. A.; Bouwman, A. F.; Mayorga, E.; Seitzinger, S. (3) Newman, E. I. Phosphorus inputs to terrestrial ecosystems. J. Magnitudes and sources of dissolved inorganic phosphorus inputs to Ecol. 1995, 83 (4), 713−726. (4) Elser, J.; Bennett, E. Phosphorus cycle: a broken biogeochemical surface fresh waters and the coastal zone: A new global model. Global − Biogeochem. Cycles 2010, 24 (1), GB1003. cycle. Nature 2011, 478 (7367), 29 31. ́ (5) Carpenter, S. R. Phosphorus control is critical to mitigating (23) Beusen, A. H.; Bouwman, A. F.; Van Beek, L. P.; Mogollon, J. eutrophication. Proc. Natl. Acad. Sci. U. S. A. 2008, 105 (32), 11039− M.; Middelburg, J. J. Global riverine N and P transport to ocean 11040. increased during the 20th century despite increased retention along − (6) Schindler, D. W.; Carpenter, S. R.; Chapra, S. C.; Hecky, R. E.; the aquatic continuum. Biogeosciences 2016, 13 (8), 2441 2451. Orihel, D. M. Reducing phosphorus to curb lake eutrophication is a (24) Bouwman, A. F.; Beusen, A. H. W.; Billen, G. Human alteration success. Environ. Sci. Technol. 2016, 50 (17), 8923−8929. of the global nitrogen and phosphorus soil balances for the period − (7) Sharpley, A.; Jarvie, H. P.; Buda, A.; May, L.; Spears, B.; 1970 2050. Global Biogeochem. Cycles 2009, 23 (4), GB0A04. Kleinman, P. Phosphorus legacy: Overcoming the effects of past (25) Bouwman, L.; Goldewijk, K. K.; Van Der Hoek, K. W.; Beusen, management practices to mitigate future water quality impairment. J. A. H. W.; Van Vuuren, D. P.; Willems, J.; Rufino, M. C.; Stehfest, E. Environ. Qual. 2013, 42 (5), 1308−1326. Exploring global changes in nitrogen and phosphorus cycles in (8) Muenich, R. L.; Kalcic, M.; Scavia, D. Evaluating the impact of agriculture induced by livestock production over the 1900−2050 legacy P and agricultural conservation practices on nutrient loads from period. Proc. Natl. Acad. Sci. U. S. A. 2013, 110 (52), 20882−20887. the Maumee River Watershed. Environ. Sci. Technol. 2016, 50 (15), (26) Chen, M.; Graedel, T. E. A half-century of global phosphorus 8146−8154. flows, stocks, production, consumption, recycling, and environmental (9) Tilman, D.; Balzer, C.; Hill, J.; Befort, B. L. Global food demand impacts. Global Environ. Change 2016, 36, 139−152. and the sustainable intensification of agriculture. Proc. Natl. Acad. Sci. (27) Sharpley, A.; Kleinman, P.; Jarvie, H.; Flaten, D. Distant views U. S. A. 2011, 108 (50), 20260−20264. and local realities: The limits of global assessments to restore the (10) Tilman, D.; Clark, M. Global diets link environmental fragmented phosphorus cycle. Agric. Environ. Lett. 2016, 1,1−5. sustainability and human health. Nature 2014, 515 (7528), 518−522. (28) Stumm, W. The acceleration of the hydrogeochemical cycling of (11) Jarvie, H. P.; Sharpley, A. N.; Flaten, D.; Kleinman, P. J. A.; phosphorus. Water Res. 1973, 7, 131−144. Jenkins, A.; Simmons, T. The pivotal role of phosphorus in a resilient (29) Lerman, A.; Mackenzie, F.; Garrels, R. Modeling of geochemical Water−Energy−Food security nexus. J. Environ. Qual. 2015, 44 (4), cycles: Phosphorus as an example. Geol. Soc. Am. Mem. 1975, 142, 1049−1062. 205−218. (12) Gruber, N.; Galloway, J. N. An Earth-system perspective of the (30) Davies, J. A. C.; Tipping, E.; Rowe, E. C.; Boyle, J. F.; Graf global . Nature 2008, 451 (7176), 293−296. Pannatier, E.; Martinsen, V. Long-term P weathering and recent N (13) Regnier, P.; Friedlingstein, P.; Ciais, P.; Mackenzie, F. T.; deposition control contemporary plant-soil C, N, and P. Global Gruber, N.; Janssens, I. A.; Laruelle, G. G.; Lauerwald, R.; Luyssaert, Biogeochem. Cycles 2016, 30 (2), 2015GB005167. S.; Andersson, A. J.; Arndt, S.; Arnosti, C.; Borges, A. V.; Dale, A. W.; (31) Graham, W. F.; Duce, R. A. Atmospheric pathways of the ́ Gallego-Sala, A.; Godderis, Y.; Goossens, N.; Hartmann, J.; Heinze, C.; phosphorus cycle. Geochim. Cosmochim. Acta 1979, 43 (8), 1195− Ilyina, T.; Joos, F.; LaRowe, D. E.; Leifeld, J.; Meysman, F. J. R.; 1208. Munhoven, G.; Raymond, P. A.; Spahni, R.; Suntharalingam, P.; (32) Filippelli, G. M. The global phosphorus cycle. Rev. Mineral. Thullner, M. Anthropogenic perturbation of the carbon fluxes from − − Geochem. 2002, 48 (1), 391 425. land to ocean. Nat. Geosci. 2013, 6 (8), 597 607. (33) Pierrou, U. The global phosphorus cycle. Ecol. Bull. 1976, 22, (14) Compton, J.; Mallinson, D.; Glenn, C.; Filippelli, G.; Foellmi, 75−88. K.; Shields, G.; Zanin, Y. Variations in the global phosphorus cycle. In (34) Richey, J. E. The phosphorus cycle. In The Major Biogeochemical Marine Authigenesis: From Global to Microbial Society for Sedimentary Cycles and Their Interactions; Bolin, B.; Cook, R. B., Eds.; John Wiley Geology; Glen, C. R.; Prevot-Lucas, L.; Lucas, J., Eds.; SEPM Special and Sons: New York, 1983; pp 51−56. Publication: OK, 2000; Vol. 66,pp21−33. (35) Yang, X.; Post, W. M.; Thornton, P. E.; Jain, A. The distribution (15) Smil, V. Phosphorus in the environment: natural flows and of soil phosphorus for global biogeochemical modeling. Biogeosciences human interferences. Annu. Rev. Energy Environ. 2000, 25 (1), 53−88. 2013, 10 (4), 2525−2537. (16) Bennett, E. M.; Carpenter, S. R.; Caraco, N. F. Human Impact (36) Jahnke, R. A. The phosphorus cycle. In Earth System Science: on Erodable Phosphorus and Eutrophication: A Global Perspective. BioScience 2001, 51 (3), 227−234. From Biogeochemical Cycles to Global Changes; Jacobson, M.; Charlson, R. J.; Rodhe, H.; Orians, G. H., Eds.; Academic Press: New York, (17) Mahowald, N.; Jickells, T. D.; Baker, A. R.; Artaxo, P.; Benitez- − Nelson, C. R.; Bergametti, G.; Bond, T. C.; Chen, Y.; Cohen, D. D.; 2000; pp 360 376. (37) Ashley, K.; Cordell, D.; Mavinic, D. A brief history of Herut, B.; Kubilay, N.; Losno, R.; Luo, C.; Maenhaut, W.; McGee, K. ’ A.; Okin, G. S.; Siefert, R. L.; Tsukuda, S. Global distribution of phosphorus: from the philosopher s stone to nutrient recovery and − atmospheric phosphorus sources, concentrations and deposition rates, reuse. Chemosphere 2011, 84 (6), 737 746. and anthropogenic impacts. Global Biogeochem. Cycles 2008, 22 (4), (38) Kelly, T. D.; Matos, G. R. Historical statistics for mineral and GB4026. material commodities in the United States (2016 version) 2014. (18) Tipping, E.; Benham, S.; Boyle, J. F.; Crow, P.; Davies, J.; http://minerals.usgs.gov/minerals/pubs/historical-statistics/ (ac- Fischer, U.; Guyatt, H.; Helliwell, R.; Jackson-Blake, L.; Lawlor, A. J.; cessed 22 December, 2016). Monteith, D. T.; Rowe, E. C.; Toberman, H. Atmospheric deposition (39) FAO FAOSTAT database. 2016.http://www.fao.org/faostat/ of phosphorus to land and freshwater. Environ. Sci.: Processes Impacts en/#home (accessed 22 December, 2016). 2014, 16 (7), 1608−1617. (40) Villalba, G.; Liu, Y.; Schroder, H.; Ayres, R. U. Global (19) Wang, R.; Balkanski, Y.; Boucher, O.; Ciais, P.; Peñuelas, J.; Tao, phosphorus flows in the industrial economy from a production S. Significant contribution of combustion-related emissions to the perspective. J. Ind. Ecol. 2008, 12 (4), 557−569. atmospheric phosphorus budget. Nat. Geosci. 2015, 8 (1), 48−54. (41) Wu, H.; Yuan, Z.; Zhang, Y.; Gao, L.; Liu, S. Life-cycle (20) Hartmann, J.; Moosdorf, N.; Lauerwald, R.; Hinderer, M.; West, phosphorus use efficiency of the farming system in Anhui Province, A. J. Global chemical weathering and associated P-releaseThe role of Central China. Resour., Conserv. Recycl. 2014, 83,1−14.

2447 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

(42) Yuan, Z.; Wu, H.; He, X.; Liu, X. A bottom−up model for case study from Busia District, Uganda. Agric., Ecosyst. Environ. 2015, quantifying anthropogenic phosphorus cycles in watersheds. J. Cleaner 207,26−39. Prod. 2014, 84 (1), 502−508. (64) Li, B.; Boiarkina, I.; Young, B.; Yu, W. Substance flow analysis of (43) Liu, Y.; Villalba, G.; Ayres, R. U.; Schroder, H. Global phosphorus within New Zealand and comparison with other countries. phosphorus flows and environmental impacts from a consumption Sci. Total Environ. 2015, 527−528, 483−492. perspective. J. Ind. Ecol. 2008, 12 (2), 229−247. (65) Smit, A. L.; van Middelkoop, J. C.; van Dijk, W.; van Reuler, H. (44) Cordell, D.; Drangert, J.-O.; White, S. The story of phosphorus: A substance flow analysis of phosphorus in the food production, Global food security and food for thought. Global Environ. Change processing and consumption system of the Netherlands. Nutr. Cycling 2009, 19 (2), 292−305. Agroecosyst. 2015, 103 (1), 1−13. (45) MacDonald, G. K.; Bennett, E. M.; Potter, P. A.; Ramankutty, (66) Prathumchai, N.; Polprasert, S.; Polprasert, C. Evaluation of N. Agronomic phosphorus imbalances across the world’s croplands. phosphorus flows in agricultural sector of Thailand. GMSARN Int. J. Proc. Natl. Acad. Sci. U. S. A. 2011, 108 (7), 3086−3091. 2016, 10, 163−170. (46) Zhang, J.; Beusen, A. H. W.; Van Apeldoorn, D. F.; Mogollon,́ J. (67) Potter, P.; Ramankutty, N.; Bennett, E. M.; Donner, S. D. M.; Yu, C.; Bouwman, A. F. Spatiotemporal dynamics of soil Characterizing the spatial patterns of global fertilizer application and phosphorus and crop uptake in global cropland during the 20th manure production. Earth Interact. 2010, 14 (2), 1−22. century. Biogeosciences 2017, 14 (8), 2055−2068. (68) Bai, Z. H.; Ma, L.; Qin, W.; Chen, Q.; Oenema, O.; Zhang, F. S. (47) Liu, X.; Sheng, H.; Jiang, S.; Yuan, Z.; Zhang, C.; Elser, J. J. Changes in pig production in China and their effects on nitrogen and Intensification of phosphorus cycling in China since the 1600s. Proc. phosphorus use and losses. Environ. Sci. Technol. 2014, 48 (21), Natl. Acad. Sci. U. S. A. 2016, 113 (10), 2609−2614. 12742−12749. (48) Antikainen, R.; Lemola, R.; Nousiainen, J. I.; Sokka, L.; Esala, (69) Robinson, T.; Thornton, P.; Franceschini, G.; Kruska, R.; M.; Huhtanen, P.; Rekolainen, S. Stocks and flows of nitrogen and Chiozza, F.; Notenbaert, A.; Cecchi, G.; Herrero, M.; Epprecht, M.; phosphorus in the Finnish food production and consumption system. Fritz, S. Global Livestock Production Systems; Food and Agriculture Agric., Ecosyst. Environ. 2005, 107 (2−3), 287−305. Organization of the United Nations (FAO): Rome, Italy/Kenay, 2011. (49) Schmid Neset, T.-S.; Bader, H.-P.; Scheidegger, R.; Lohm, U. (70) Sheldrick, W.; Keith Syers, J.; Lingard, J. Contribution of The flow of phosphorus in food production and consumption  livestock excreta to nutrient balances. Nutr. Cycling Agroecosyst. 2003, Linköping, Sweden, 1870−2000. Sci. Total Environ. 2008, 396 (2−3), 66 (2), 119−131. 111−120. (71) Carpenter, S. R.; Caraco, N. F.; Correll, D. L.; Howarth, R. W.; (50) Jeong, Y.-S.; Matsubae-Yokoyama, K.; Kubo, H.; Pak, J.-J.; Sharpley, A. N.; Smith, V. H. Nonpoint pollution of surface waters Nagasaka, T. Substance flow analysis of phosphorus and manganese with phosphorus and nitrogen. Ecol. Appl. 1998, 8 (3), 559−568. correlated with South Korean steel industry. Resour., Conserv. Recycl. (72) Institute of Medicine. Dietary Reference Intakes for Calcium, 2009, 53 (9), 479−489. Phosphorus, Magnesium, Vitamin D, and Fluoride; National Academy (51) Matsubae-Yokoyama, K.; Kubo, H.; Nakajima, K.; Nagasaka, T. Press: Washington, DC, 1997. A material flow analysis of phosphorus in Japan. J. Ind. Ecol. 2009, 13 (73) Mihelcic, J. R.; Fry, L. M.; Shaw, R. Global potential of (5), 687−705. phosphorus recovery from human urine and feces. Chemosphere 2011, (52) Seyhan, D. Country-scale phosphorus balancing as a base for 84 (6), 832−839. resources conservation. Resour., Conserv. Recy. 2009, 53 (12), 698− (74) Yu, X. The treatment of night soil and waste in Modern China. 709. In Health and Hygiene in Chinese East Asia Politics and Publics in the (53) Ghani, L.; Mahmood, N. Balance sheet for phosphorus in Long Twentieth Century; Ki, A.; Leung, C.; Furth, C., Eds. Duke Malaysia by SFA. Aust. J. Basic Appl. Sci. 2011, 5 (12), 3069−3079. University Press: Durham and London, 2010; pp 51−72. (54) Matsubae, K.; Kajiyama, J.; Hiraki, T.; Nagasaka, T. Virtual (75) Chen, Z.; Tang, Y. Study on sustainable use of urban night-soil phosphorus ore requirement of Japanese economy. Chemosphere 2011, in China. Urban Environ. Urban Ecol. 1999, 12 (2), 42−49. 84 (6), 767−772. (76) Gu, B.; Leach, A. M.; Ma, L.; Galloway, J. N.; Chang, S. X.; Ge, (55) Suh, S.; Yee, S. Phosphorus use-efficiency of agriculture and Y.; Chang, J. Nitrogen footprint in China: food, energy, and nonfood food system in the US. Chemosphere 2011, 84 (6), 806−813. goods. Environ. Sci. Technol. 2013, 47 (16), 9217−9224. (56) Linderholm, K.; Mattsson, J. E.; Tillman, A.-M. Phosphorus (77) Yuan, Z.; Shi, J.; Wu, H.; Zhang, L.; Bi, J. Understanding the flows to and from Swedish agriculture and food chain. Ambio 2012, 41 anthropogenic phosphorus pathway with substance flow analysis at the (8), 883−893. city level. J. Environ. Manage. 2011, 92 (8), 2021−2028. (57) MacDonald, G. K.; Bennett, E. M.; Carpenter, S. R. Embodied (78) Yuan, Z.; Sun, L.; Bi, J.; Wu, H.; Zhang, L. Phosphorus flow phosphorus and the global connections of United States agriculture. analysis of the socioeconomic ecosystem of Shucheng County, China. Environ. Res. Lett. 2012, 7 (4), 044024. Ecol. Appl. 2011, 21 (7), 2822−2832. (58) Senthilkumar, K.; Nesme, T.; Mollier, A.; Pellerin, S. (79) Quinton, J. N.; Govers, G.; Van Oost, K.; Bardgett, R. D. The Conceptual design and quantification of phosphorus flows and impact of agricultural soil erosion on biogeochemical cycling. Nat. balances at the country scale: The case of France. Global Biogeochem. Geosci. 2010, 3 (5), 311−314. Cycles 2012, 26 (2), GB2008. (80) Lwin, C. M.; Murakami, M.; Hashimoto, S. The implications of (59) Cooper, J.; Carliell-Marquet, C. A substance flow analysis of allocation scenarios for global phosphorus flow from agriculture and phosphorus in the UK food production and consumption system. wastewater. Resour., Conserv. Recycl. 2017, 122 (Supplement C), 94− Resour., Conserv. Recy. 2013, 74,82−100. 105. (60) Cordell, D.; Jackson, M.; White, S. Phosphorus flows through (81) García-Ruiz, J. M.; Beguería, S.; Nadal-Romero, E.; Gonzalez-́ the Australian food system: Identifying intervention points as a Hidalgo, J. C.; Lana-Renault, N.; Sanjuan,́ Y. A meta-analysis of soil roadmap to phosphorus security. Environ. Sci. Policy 2013, 29,87−102. erosion rates across the world. Geomorphology 2015, 239, 160−173. (61) Egle, L.; Zoboli, O.; Thaler, S.; Rechberger, H.; Zessner, M. The (82) Keesstra, S.; Pereira, P.; Novara, A.; Brevik, E. C.; Azorin- Austrian P budget as a basis for resource optimization. Resour., Conserv. Molina, C.; Parras-Alcantara,́ L.; Jordan,́ A.; Cerda,̀ A. Effects of soil Recy. 2014, 83, 152−162. management techniques on soil water erosion in apricot orchards. Sci. (62) Jedelhauser, M.; Binder, C. R. Losses and efficiencies of Total Environ. 2016, 551−552, 357−366. phosphorus on a national level - A comparison of European substance (83) Pimentel, D.; Harvey, C.; Resosudarmo, P.; Sinclair, K.; Kurz, flow analyses. Resour., Conserv. Recycl. 2015, 105, 294−310. D.; McNair, M.; Crist, S.; Shpritz, L.; Fitton, L.; Saffouri, R.; Blair, R. (63) Lederer, J.; Karungi, J.; Ogwang, F. The potential of wastes to Environmental and Economic Costs of Soil Erosion and Conservation improve nutrient levels in agricultural soils: A material flow analysis Benefits. Science 1995, 267 (5201), 1117−1123.

2448 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

(84) Meybeck, M. Carbon, nitrogen, and phosphorus transport by (104) IFA IFADATA. 2016. http://ifadata.fertilizer.org/ucSearch. world rivers. Am. J. Sci. 1982, 282 (4), 401−450. aspx (accessed 22 December, 2016). (85) Seitzinger, S. P.; Mayorga, E.; Bouwman, A. F.; Kroeze, C.; (105) Cooper, J.; Lombardi, R.; Boardman, D.; Carliell-Marquet, C. Beusen, A. H. W.; Billen, G.; Van Drecht, G.; Dumont, E.; Fekete, B. The future distribution and production of global phosphate rock M.; Garnier, J.; Harrison, J. A. Global river nutrient export: A scenario reserves. Resour., Conserv. Recycl. 2011, 57,78−86. analysis of past and future trends. Global Biogeochem. Cycles 2010, 24 (106) UN. Prospects: The 2015 Revision, Volume I: (4), GB0A08. Comprehensive Tables. United Nations (UN), Department of (86) Mayorga, E.; Seitzinger, S. P.; Harrison, J. A.; Dumont, E.; Economic and Social Affairs, Population Division, New York, 2016. Beusen, A. H. W.; Bouwman, A. F.; Fekete, B. M.; Kroeze, C.; Van (107) Metson, G. S.; Bennett, E. M.; Elser, J. J. The role of diet in Drecht, G. Global nutrient export from watersheds 2 (NEWS 2): phosphorus demand. Environ. Res. Lett. 2012, 7 (4), 044043. Model development and implementation. Environ. Modell. Software (108) Vitousek, P. M.; Naylor, R.; Crews, T.; David, M. B.; 2010, 25 (7), 837−853. Drinkwater, L. E.; Holland, E.; Johnes, P. J.; Katzenberger, J.; (87) Jiang, S.; Yuan, Z. Phosphorus flow patterns in the Chaohu Martinelli, L. A.; Matson, P. A.; Nziguheba, G.; Ojima, D.; Palm, C. A.; watershed from 1978 to 2012. Environ. Sci. Technol. 2015, 49 (24), Robertson, G. P.; Sanchez, P. A.; Townsend, A. R.; Zhang, F. S. 13973−13982. Nutrient Imbalances in Agricultural Development. Science 2009, 324 (88) Sattari, S. Z.; Bouwman, A. F.; Giller, K. E.; van Ittersum, M. K. (5934), 1519−1520. Residual soil phosphorus as the missing piece in the global phosphorus (109) Peñuelas, J.; Poulter, B.; Sardans, J.; Ciais, P.; van der Velde, crisis puzzle. Proc. Natl. Acad. Sci. U. S. A. 2012, 109 (16), 6348−6353. M.; Bopp, L.; Boucher, O.; Godderis, Y.; Hinsinger, P.; Llusia, J.; (89) Liu, J.; Hu, Y.; Yang, J.; Abdi, D.; Cade-Menun, B. J. Nardin, E.; Vicca, S.; Obersteiner, M.; Janssens, I. A. Human-induced Investigation of soil legacy phosphorus transformation in long-term nitrogen−phosphorus imbalances alter natural and managed ecosys- agricultural fields using sequential fractionation, P K-edge XANES and tems across the globe. Nat. Commun. 2013, 4, 2394. solution P-NMR spectroscopy. Environ. Sci. Technol. 2015, 49 (1), (110) Fink, G.; Flörke, M.; Alcamo, J. Global baseline data on 168−176. phosphorus pollution of large lakes. In EGU General Assembly (90) Verloop, J.; Oenema, J.; Burgers, S.; Aarts, H.; Van Keulen, H. Conference 2016; Vol. 18, p 11731. P-equilibrium fertilization in an intensive dairy farming system: Effects (111) CIESIN. Gridded Population of the World, Version 4 on soil-P status, crop yield and P leaching. Nutr. Cycling Agroecosyst. (GPWv4): Population Count Grid. 2015. http://sedac.ciesin. 2010, 87 (3), 369−382. columbia.edu/data/collection/gpw-v4/sets/browse (accessed 22 De- (91) Mishima, S.; Endo, A.; Kohyama, K. Recent trends in phosphate cember, 2016). balance nationally and by region in Japan. Nutr. Cycling Agroecosyst. (112) Wu, P.; Qin, B.; Yu, G. Estimates of long-term water total 2010, 86 (1), 69−77. phosphorus (TP) concentrations in three large shallow lakes in the (92) Myriokefalitakis, S.; Nenes, A.; Baker, A. R.; Mihalopoulos, N.; Yangtze River basin, China. Environ. Sci. Pollut. Res. 2016, 23 (5), Kanakidou, M. Bioavailable atmospheric phosphorous supply to the 4938−4948. global ocean: A 3-D global modeling study. Biogeosciences 2016, 13 (113) Jarvie, H. P.; Sharpley, A. N.; Spears, B.; Buda, A. R.; May, L.; (24), 6519−6543. Kleinman, P. J. Water quality remediation faces unprecedented (93) Okin, G. S.; Mahowald, N.; Chadwick, O. A.; Artaxo, P. Impact challenges from “legacy phosphorus. Environ. Sci. Technol. 2013, 47 of desert dust on the of phosphorus in terrestrial (16), 8997−8998. ecosystems. Global Biogeochem. Cycles 2004, 18 (2), GB2005. (114) Jarvie, H. P.; Johnson, L. T.; Sharpley, A. N.; Smith, D. R.; (94) Neff, J. C.; Ballantyne, A. P.; Farmer, G. L.; Mahowald, N. M.; Baker, D. B.; Bruulsema, T. W.; Confesor, R. Increased soluble Conroy, J. L.; Landry, C. C.; Overpeck, J. T.; Painter, T. H.; Lawrence, phosphorus loads to Lake Erie: Unintended consequences of C. R.; Reynolds, R. L. Increasing eolian dust deposition in the western conservation practices? J. Environ. Qual. 2017, 46 (1), 123−132. United States linked to human activity. Nat. Geosci. 2008, 1 (3), 189− (115) Powers, S. M.; Bruulsema, T. W.; Burt, T. P.; Chan, N. I.; Elser, 195. J. J.; Haygarth, P. M.; Howden, N. J.; Jarvie, H. P.; Lyu, Y.; Peterson, (95) Fixen, P. E. World fertilizer nutrient reserves−a view to the H. M. Long-term accumulation and transport of anthropogenic future. Better Crops 2009, 93 (3), 8−11. phosphorus in three river basins. Nat. Geosci. 2016, 9 (5), 353−356. (96) Van Vuuren, D. P.; Bouwman, A. F.; Beusen, A. H. W. (116) Michalak, A. M.; Anderson, E. J.; Beletsky, D.; Boland, S.; Phosphorus demand for the 1970−2100 period: A scenario analysis of Bosch, N. S.; Bridgeman, T. B.; Chaffin, J. D.; Cho, K.; Confesor, R.; . Global Environ. Change 2010, 20 (3), 428−439. Daloglu,̆ I.; DePinto, J. V.; Evans, M. A.; Fahnenstiel, G. L.; He, L.; Ho, (97) Smit, A. L.; Bindraban, P. S.; Schröder, J.; Conijn, J.; Van der J. C.; Jenkins, L.; Johengen, T. H.; Kuo, K. C.; LaPorte, E.; Liu, X.; Meer, H. Phosphorus in Agriculture: Global Resources, Trends and McWilliams, M. R.; Moore, M. R.; Posselt, D. J.; Richards, R. P.; Developments; Plant Research International, Wageningen University: Scavia, D.; Steiner, A. L.; Verhamme, E.; Wright, D. M.; Zagorski, M. Wageningen, Netherlands, 2009. A. Record-setting in Lake Erie caused by agricultural and (98) Vaccari, D. A. Phosphorus: a looming crisis. Sci. Am. 2009, 300 meteorological trends consistent with expected future conditions. Proc. (6), 54−59. Natl. Acad. Sci. U. S. A. 2013, 110 (16), 6448−6452. (99) USGS Mineral Commodity Summaries 2010. 2010. https:// (117) Smith, V. H.; Joye, S. B.; Howarth, R. W. Eutrophication of minerals.usgs.gov/minerals/pubs/mcs/ (accessed 22 December, freshwater and marine ecosystems. Limnol. Oceanogr. 2006, 51 2016). (1part2), 351−355. (100) USGS Mineral Commodity Summaries 2011. 2011. https:// (118) Schindler, D. W.; Hecky, R. E.; Findlay, D. L.; Stainton, M. P.; minerals.usgs.gov/minerals/pubs/mcs/ (accessed 22 December, Parker, B. R.; Paterson, M. J.; Beaty, K. G.; Lyng, M.; Kasian, S. E. 2016). Eutrophication of lakes cannot be controlled by reducing nitrogen (101) Koppelaar, R. H. E. M.; Weikard, H. P. Assessing phosphate input: Results of a 37-year whole-ecosystem experiment. Proc. Natl. rock depletion and phosphorus recycling options. Global Environ. Acad. Sci. U. S. A. 2008, 105 (32), 11254−11258. Change 2013, 23 (6), 1454−1466. (119) Howarth, R. W.; Marino, R. Nitrogen as the limiting nutrient (102) Van Kauwenbergh, S. J. World Phosphate Rock Reserves and for eutrophication in coastal marine ecosystems: evolving views over Resources; International Fertilizer Development Centre (IFDC): WA, three decades. Limnol. Oceanogr. 2006, 51 (1part2), 364−376. 2010. (120) Yang, X.; Wu, X.; Hao, H.; He, Z. Mechanisms and assessment (103) USGS Mineral Commodity Summaries 2016. 2016. https:// of water eutrophication. J. Zhejiang Univ., Sci., B 2008, 9 (3), 197−209. minerals.usgs.gov/minerals/pubs/mcs/ (accessed 22 December, (121) Azevedo, L. B.; van Zelm, R.; Elshout, P. M. F.; Hendriks, A. J.; 2016). Leuven, R. S. E. W.; Struijs, J.; de Zwart, D.; Huijbregts, M. A. J.

2449 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450 Environmental Science & Technology Critical Review

Species richness−phosphorus relationships for lakes and streams (139) Wang, R.; Goll, D.; Balkanski, Y.; Hauglustaine, D.; Boucher, worldwide. Global Ecol. Biogeogr. 2013, 22 (12), 1304−1314. O.; Ciais, P.; Janssens, I.; Penuelas, J.; Guenet, B.; Sardans, J.; Bopp, L.; (122) Azevedo, L. B.; Henderson, A. D.; van Zelm, R.; Jolliet, O.; Vuichard, N.; Zhou, F.; Li, B.; Piao, S.; Peng, S.; Huang, Y.; Tao, S. Huijbregts, M. A. J. Assessing the importance of spatial variability Global forest carbon uptake due to nitrogen and phosphorus versus model choices in life cycle impact assessment: The case of deposition from 1850 to 2100. Global Change Biol. 2017, 23 (11), freshwater eutrophication in Europe. Environ. Sci. Technol. 2013, 47 4854−4872. (23), 13565−13570. (140) Ma, L.; Velthof, G. L.; Wang, F. H.; Qin, W.; Zhang, W. F.; (123) Scherer, L.; Pfister, S. Modelling spatially explicit impacts from Liu, Z.; Zhang, Y.; Wei, J.; Lesschen, J. P.; Ma, W. Q.; Oenema, O.; phosphorus emissions in agriculture. Int. J. Life Cycle Assess. 2015, 20 Zhang, F. S. Nitrogen and phosphorus use efficiencies and losses in the (6), 785−795. food chain in China at regional scales in 1980 and 2005. Sci. Total − (124) Qin, B.; Zhu, G.; Gao, G.; Zhang, Y.; Li, W.; Paerl, H. W.; Environ. 2012, 434,51 61. ̃ Carmichael, W. W. A drinking water crisis in Lake Taihu, China: (141) Wang, W.; Sardans, J.; Zeng, C.; Zhong, C.; Li, Y.; Penuelas, J. Linkage to climatic variability and lake management. Environ. Manage. Responses of soil nutrient concentrations and stoichiometry to − different human land uses in a subtropical tidal wetland. Geoderma 2010, 45 (1), 105 112. − − (125) Jetoo, S.; Grover, V.; Krantzberg, G. The Toledo drinking 2014, 232 234, 459 470. water advisory: Suggested application of the water safety planning (142) Liu, H.; Zhou, J.; Shen, J.; Li, Y.; Li, Y.; Ge, T.; Guggenberger, approach. Sustainability 2015, 7 (8), 9787−9808. G.; Wu, J. Agricultural uses reshape soil C, N, and P stoichiometry in ̈ ̈ subtropical ecosystems. Biogeosciences 2016, 2016,1−23. (126) Muller, B.; Bryant, L. D.; Matzinger, A.; Wuest, A. ̃ Hypolimnetic oxygen depletion in eutrophic lakes. Environ. Sci. (143) Yan, Z.; Han, W.; Penuelas, J.; Sardans, J.; Elser, J. J.; Du, E.; Technol. 2012, 46 (18), 9964−9971. Reich, P. B.; Fang, J. Phosphorus accumulates faster than nitrogen (127) Jenny, J.-P.; Francus, P.; Normandeau, A.; Lapointe, F.; Perga, globally in freshwater ecosystems under anthropogenic impacts. Ecol. Lett. 2016, 19 (10), 1237−1246. M.-E.; Ojala, A.; Schimmelmann, A.; Zolitschka, B. Global spread of (144) Peñuelas, J.; Sardans, J.; Rivas-ubach, A.; Janssens, I. A. The hypoxia in freshwater ecosystems during the last three centuries is human-induced imbalance between C, N and P in Earth’s life system. caused by rising local human pressure. Global Change Biol. 2016, 22 − − Global Change Biol. 2012, 18 (1), 3 6. (4), 1481 1489. (145) Reed, S. C.; Yang, X.; Thornton, P. E. Incorporating (128) Vonlanthen, P.; Bittner, D.; Hudson, A. G.; Young, K. A.; phosphorus cycling into global modeling efforts: a worthwhile, Muller, R.; Lundsgaard-Hansen, B.; Roy, D.; Di Piazza, S.; Largiader, tractable endeavor. New Phytol. 2015, 208 (2), 324−329. C. R.; Seehausen, O. Eutrophication causes speciation reversal in − (146) Sun, Y.; Peng, S.; Goll, D. S.; Ciais, P.; Guenet, B.; whitefish adaptive radiations. Nature 2012, 482 (7385), 357 362. Guimberteau, M.; Hinsinger, P.; Janssens, I. A.; Peñuelas, J.; Piao, (129) Goll, D.; Brovkin, V.; Parida, B.; Reick, C. H.; Kattge, J.; Reich, ̈ S.; Poulter, B.; Violette, A.; Yang, X.; Yin, Y.; Zeng, H. Diagnosing P.; Van Bodegom, P.; Niinemets, U. Nutrient limitation reduces land phosphorus limitations in natural terrestrial ecosystems in carbon uptake in simulations with a model of combined carbon, models. Earth's Future 2017, 5 (7), 730−749. − nitrogen and phosphorus cycling. Biogeosciences 2012, 9 (9), 3547 (147) Hartmann, J.; Lauerwald, R.; Moosdorf, N. A brief overview of 3569. the GLObal RIver Chemistry Database, GLORICH. Procedia Earth (130) Cleveland, C. C.; Townsend, A. R.; Schimel, D. S.; Fisher, H.; Planet. Sci. 2014, 10,23−27. Howarth, R. W.; Hedin, L. O.; Perakis, S. S.; Latty, E. F.; Von Fischer, (148) Maavara, T.; Parsons, C. T.; Ridenour, C.; Stojanovic, S.; Dürr, J. C.; Elseroad, A.; Wasson, M. F. Global patterns of terrestrial H. H.; Powley, H. R.; Van Cappellen, P. Global phosphorus retention biological nitrogen (N2) fixation in natural ecosystems. Global by river damming. Proc. Natl. Acad. Sci. U. S. A. 2015, 112 (51), Biogeochem. Cycles 1999, 13 (2), 623−645. 15603−15608. (131) Vitousek, P. M.; Porder, S.; Houlton, B. Z.; Chadwick, O. A. (149) Kebreab, E.; Hansen, A. V.; Strathe, A. B. Animal production Terrestrial phosphorus limitation: mechanisms, implications, and for efficient phosphate utilization: from optimized feed to high nitrogen−phosphorus interactions. Ecol. Appl. 2010, 20 (1), 5−15. efficiency livestock. Curr. Opin. Biotechnol. 2012, 23 (6), 872−877. (132) Elser, J. J.; Bracken, M. E.; Cleland, E. E.; Gruner, D. S.; (150) Li, H.; Huang, G.; Meng, Q.; Ma, L.; Yuan, L.; Wang, F.; Harpole, W. S.; Hillebrand, H.; Ngai, J. T.; Seabloom, E. W.; Shurin, J. Zhang, W.; Cui, Z.; Shen, J.; Chen, X.; Jiang, R.; Zhang, F. Integrated B.; Smith, J. E. Global analysis of nitrogen and phosphorus limitation soil and plant phosphorus management for crop and environment in of primary producers in freshwater, marine and terrestrial ecosystems. China. A review. Plant Soil 2011, 349 (1−2), 157−167. Ecol. Lett. 2007, 10 (12), 1135−1142. (133) Zheng, M.; Li, D.; Lu, X.; Zhu, X.; Zhang, W.; Huang, J.; Fu, S.; Lu, X.; Mo, J. Effects of phosphorus addition with and without nitrogen addition on biological in tropical legume and non-legume tree plantations. Biogeochemistry 2016, 131 (1), 65−76. (134) Mao, R.; Chen, H.-M.; Zhang, X.-H.; Shi, F.-X.; Song, C.-C. Effects of P addition on plant C:N:P stoichiometry in an N-limited temperate wetland of Northeast China. Sci. Total Environ. 2016, 559, 1−6. (135) Mao, R.; Zeng, D.-H.; Zhang, X.-H.; Song, C.-C. Responses of plant nutrient resorption to phosphorus addition in freshwater marsh of Northeast China. Sci. Rep. 2015, 5, 8097. (136) Mao, R.; Li, S.-Y.; Zhang, X.-H.; Wang, X.-W.; Song, C.-C. Effect of long-term phosphorus addition on the quantity and quality of dissolved organic carbon in a freshwater wetland of Northeast China. Sci. Total Environ. 2017, 586, 1032−1037. (137) Peng, X.; Peng, Y.; Yue, K.; Deng, Y. Different responses of terrestrial C, N, and P pools and C/N/P ratios to P, NP, and NPK addition: A meta-analysis. Water, Air, Soil Pollut. 2017, 228 (6), 197. (138) Li, Y.; Niu, S.; Yu, G. Aggravated phosphorus limitation on biomass production under increasing nitrogen loading: a meta-analysis. Global Change Biol. 2016, 22 (2), 934−943.

2450 DOI: 10.1021/acs.est.7b03910 Environ. Sci. Technol. 2018, 52, 2438−2450