Iowa’s Annual Phosphorus Budget: Quantifying the Inputs and Outputs of Phosphorus Transport Processes

Abstract

Phosphorus (P) plays an integral part in Iowa’s economic and environmental activities through its role as an essential nutrient and waterborne pollutant. However, the amount of phosphorus transported through these activities has not been well quantified. This study estimates the annual mass of P entering and exiting Iowa’s landscape from 1998 to 2022 through seven transport pathways. Four input pathways (fertilizer application, manure production, industrial sources, and human sources) and three output pathways (harvesting, livestock grazing, and stream export) were quantified using various agricultural, economic, and water quality datasets. We also estimated the total mass of P present in the top 0.61 m layer of Iowa’s landscape using results from a statewide soil sampling survey. The harvest component was the largest, with annual values consistently above 200 million kg. This was followed by the fertilizer and manure components, with annual values near 100 million kg. The other components were much smaller; the mean grazing and stream export values were 15 and 19 million kg, respectively, and human and industrial sources were less than 4 million kg. Stream export was the most dynamic pathway, with the largest coefficient of variation (0.59). The net P budget (inputs–outputs) was negative in 20 of the 25 years assessed, indicating that Iowa typically runs a P deficit. A trend analysis revealed that the manure, human, industry, and harvesting components increased across the 1998–2022 period while the grazing component decreased. The mass of P in Iowa’s top layer of soil was 81.5 billion kg—orders of magnitude larger than any individual budget component. This analysis provides a new perspective on P transport pathways in Iowa and may help inform policymakers as they make decisions on the many activities involving P.

1. Introduction

Phosphorus (P) is an essential macronutrient required by all plant and animal lifeforms but is often considered a pollutant after entering surface water. Excess P is a major contributor to the development of hypoxic conditions in the Gulf of Mexico [1] and around the world [2] and is viewed by researchers as the limiting nutrient in phytoplankton growth in freshwater ecosystems and some coastal waters [3,4]. The economic impact of excessive P is evident in many agricultural [5,6], municipal [7], and industrial sectors [8]. In the Midwestern U.S. state of Iowa, where P is inseparably linked to the agricultural economy and environment, researchers have investigated P’s presence in various transport pathways [9,10,11,12]. Efforts have focused on quantifying stream export [9,13], and stakeholders are investing significant resources into conservation practices that reduce P loss [14]. Monitoring riverine P loads can reveal which management practices and locations are the most useful to focus on in achieving P reduction goals [15].

P transport is of great importance globally and locally within Iowa. As with many chemicals, P travels throughout Earth as part of a biogeochemical cycle—most prominently in the lithosphere, hydrosphere, and biosphere [16]. Anthropogenic activities, such as agricultural development, urbanization, and industrialization, have drastically altered P’s natural biogeochemical cycle [17], and P now routinely exceeds its natural levels in various parts of the world [18]. This shift can be attributed to P’s role as a nutrient, mainly through its mining and application as a fertilizer. P is also a finite resource, and the mines where P is extracted have limited quantities [19]. When fertilizers are applied inefficiently, P ends up in aquatic ecosystems where it is difficult to recover [20]. Researchers have coined the term peak P, where humanity achieves its maximum rate of P production, to define this concept [21]. P will become more challenging to obtain and utilize as fertilizer past this threshold, thereby jeopardizing the world’s food security [22].

P also presents a host of environmental concerns, as it becomes a pollutant when it enters surface water. P is the limiting nutrient in most freshwater ecosystems [3]. When elevated P levels are present, certain organisms can exploit the altered water quality to the detriment of others, driving eutrophication [18]. In many instances, even small amounts of P can impair a waterbody [23], and actions that prevent P from reaching surface waters greatly reduce the likelihood and magnitude of these impairments [24]. The challenges surrounding P have made it one of the most closely studied pollutants in recent decades; concerns are especially prominent in Iowa, which contains one of the world’s most heavily altered landscapes [25,26].

A P budget can be useful for quantifying the flux of all P forms in and out of a region’s landscape [27]. It essentially estimates the mass balance of P within a region’s soils and surface water. However, the specific framework of the P budget will be dependent upon the nutrient’s biogeochemical cycle [28] that accounts for all the environmental processes whereby P can enter or exit the landscape. A P budget was previously created for Iowa [29] based on a well-known template devised by the Committee on Environment and Natural Resources [30]. The budget template used seven components—four inputs (fertilizer application, manure production, industrial sources, and human sources) and three outputs (grain harvest, livestock grazing, and stream export)—to provide a large-scale P mass balance for the state. The Iowa P budget was created to represent a typical year in the early 2000s by focusing solely on the period of 1997–2002 [29]. While this study proved to be a valuable first step in describing P processes in Iowa, it had several limitations. Primarily, it did not contain a temporal component, and stakeholders were thus unable to identify potential trends or quantify year-to-year variations in P transport. Likewise, it was not possible to explore relationships between budget components or evaluate the impact of changes in P mass balance on soil P storage in Iowa’s landscape.

Most nutrient budgets, including Iowa’s original P budget, rely on data obtained through surveys or census programs designed to collect demographic or economic information [31]. Data from these programs are essential in estimating P’s transport. Most pertinent to Iowa is the Census of Agriculture, which is conducted by the United States Department of Agriculture every five years. This program quantifies numerous agricultural activities in Iowa, such as land use areas, crop harvests, and livestock totals [32]. Annual estimates are also generated using the results of the five-year questionnaires [33]. Various state agencies in Iowa also track P movement through fertilizer distribution reports [34] and permitting point source discharges [35]. Budgets also depend upon sampling data—empirical measurements of P in the environment. Most relevant are water quality and soil samples, which track P’s presence in the hydrosphere and lithosphere [36].

In this study, we calculated an annual P budget for the State of Iowa for the 1998 to 2022 period that accounts for the significant factors that drive P storage and transport in Iowa’s environment. Using the previous budget as a guide, we calculated an annual P budget that reflected year-to-year variability of budget elements over the past two decades that accounts for changing agriculture patterns in Iowa toward more row crop production and concentrated animal storage along with increasing weather variability [37]. Further, we added additional context for the P budget by evaluating how the changing net P balance impacts P storage in Iowa soils. Overall, the goals of the study were to (1) identify magnitudes and trends in P budget components, (2) examine potential relationships among the various components, and (3) use the net P balance to assess the long-term sustainability of P in Iowa’s soils.

2. Materials and Methods

The framework of our budget involves quantifying the yearly P mass for each of the seven budget components. These components characterize P’s transport into and out of Iowa’s landscape and include four inputs (fertilizer application, manure production, industrial sources, and human sources) and three outputs (harvesting, livestock grazing, and stream export). These estimates produce a 25-year time series for each input and output.

Our annual P budget for Iowa began in 1998, the most recent year in which water sampling occurred at a statewide level. Input and output data were aggregated at an annual scale based on standard methods and widely available datasets [29]. P may exist in various chemical forms, but only the elemental form of P was tabulated in the final estimates; the masses of other elements bound to P were not included in the budget. A summary of inputs and outputs and their corresponding methodologies are provided below.

2.1. Inputs

As noted in previous budget descriptions [29,38,39], estimating these components primarily consists of converting agricultural and census data into comparative P masses.

2.1.1. Fertilizer Application

The fertilizer component accounts for the P applied to crops as a macronutrient that facilitates plant growth. Within Iowa, all data on fertilizer sales and distribution are tracked by the Iowa Department of Agriculture and Land Stewardship (IDALS) [40]. This agency documents the amount of various types of fertilizer applied. Fertilizer usage is described every year and is further broken down into the first and second halves of the crop year, July–December and January–June. The masses of individual macronutrients (P, nitrogen, and potassium) were determined by multiplying the nutrient percentage of each product by the total amount of the product sold. These P masses were then aggregated on an annual basis.

2.1.2. Manure Production

The manure component accounts for the P in animal waste generated by livestock production. Animal populations were obtained from the National Agricultural Statistics Service (NASS), a data access tool operated by the United States Department of Agriculture (USDA), which documents livestock operations and headcounts throughout the country [41]. This database provided reliable inventories for Iowa’s livestock operations. Cattle, sheep, swine, chicken, and turkey are the primary livestock animals in Iowa, and the NASS database provides the number of these animals raised each year.

P content of each species’ excrement was obtained from ISU Technical Document MWPS-18 [42]. Quantifying this component involves disaggregating livestock totals into their most basic categories (i.e., ages and sub-types) to apply the appropriate manure factor [43]. Multiplying the factors provided in Document MWPS-18 by yearly livestock totals yields the P manure inputs for each species.

2.1.3. Industrial Sources

Industrial sources of P primarily stem from wastewater effluent discharged by factories throughout Iowa. This metric is difficult to quantify directly, as there is little observed data that documents the specific nutrient content of industrial plants found along Iowa’s major rivers. However, studies that tried to measure industrial nutrient pollution have found that it correlates strongly to population [29,44]. More people present within a location leads to more industrial activities. When viewed across a large spatial region, such as the Midwest, these industrial activities emit a mostly constant percentage of P in their effluent [45]. While specific plants will emit more P than others, the differences tend to even out at the statewide level [39].

By making these assumptions, P from industrial sources can be estimated on a per capita basis. Hypoxia Task Force studies have determined that a region’s industrial discharge of P tends to be approximately 0.2 kg of P per person per year [46,47]. Multiplying this factor by Iowa’s population yields a rudimentary estimate of the industrial components of the budget.

2.1.4. Human Sources

The human component accounts for P generated in the excrement of Iowa’s residents. This feature is calculated in much the same way as the industrial component, as it can be estimated based on the population of Iowa. Studies have identified that the average person produces 1.09 kg of P per year [46,47]. This factor can be multiplied by Iowa’s annual population to estimate the human P input.

2.2. Outputs

P output was estimated over the same timeframe as the P inputs (1998–2022) and considered crop harvesting, livestock grazing, and stream export.

2.2.1. Harvesting

The harvesting component contains all the P removed from Iowa’s landscape when the state’s crops are harvested. This was estimated by tracking the number of each type of crop produced in the state. These data are available within the NASS database [41]. Corn, soybeans, oats, wheat, and hay are the major contributors to harvested P. There are additional crops harvested in Iowa beyond those described here, but their total production is small relative to the others and not likely to be a significant contributor to the overall budget [29,48].

Agricultural researchers have investigated the specific nutrient content contained in common crops. ISU Technical Document PM1688 provides these estimates for the common crops seen in Iowa [49]. The values from this document represent the average content of each nutrient present per unit of harvest. Multiplying the estimated P content of a crop by the yearly crop total yields the P harvest outputs for each commodity.

2.2.2. Livestock Grazing

The livestock grazing component represents the P directly consumed by livestock on Iowa’s pastureland. The grazing component is not directly tabulated by any organization. Rather, calculating the grazing P component requires taking the yearly area of pastureland in production and multiplying it by that year’s average hay yield.

As part of a USDA agricultural census, specific pastureland values are only tabulated once every five years. Yearly pastureland areas were interpolated from these five-year values. Non-alfalfa hay is the typical crop fed to animals on pasture [29]. This commodity is tabulated annually and represents the amount of non-alfalfa hay produced per acre. Multiplying the pastureland by the non-alfalfa hay yield results in the total amount of hay grown on Iowa’s pastures each year. Again, the NASS database provides both of these required datasets [41].

Once the total amount of pastureland hay was calculated, the P content can be found using the values provided in Document PM1688 [49]. This document gives the average P content within a typical ton of non-alfalfa hay. Multiplying this factor by the total amount of hay results in the yearly grazing component of the P budget.

2.2.3. Stream Export

P is exported from Iowa in river water flux. There are 16 locations near Iowa’s border where waterborne P has been measured routinely since 1998 (Figure 1). These sites represent the farthest downstream points along Iowa’s major rivers where water quality P data have been collected. All of Iowa’s prominent waterbodies and over 90% of its area are captured within the watersheds of these 16 locales. Every site also has a co-located United States Geological Survey (USGS) stream gauge that monitors streamflow. Each stream gauge has continuously operated since 1998, resulting in robust, uninterrupted records of daily streamflow. Previous nutrient estimates have summed loads from these 16 locations to generate statewide totals [37].

Waterborne P is measured monthly at each of the 16 sites through the Iowa Department of Natural Resources (IDNR) ambient monitoring program [50]. The USGS likewise conducted monthly P sampling from 2004 to 2014 at nine of the sites through the Big River Study [51]. These monthly samples include the concentration of P in all its soluble and particulate forms—often referred to as Total P. Both datasets are publicly available in the EPA STORET databases and were retrieved for this study. Consequently, every location has a consistent monthly record containing over 250 P samples spanning 1998 to the present. The number of P samples collected at each site is listed in

Table 1. Iowa’s 16 terminal monitoring sites where stream P loads were modeled.

Site Number	Short Name	Full Name	IDNRid	USGSid	P Samples	Area (km2)	Lat	Long
1	Boyer	Boyer River at Logan, IA	10430001	06609500	397	2256	41.64169	−95.7823
2	Des Moines	Des Moines River at Keosauqua, IA	10890001	05490500	498	36,358	40.72781	−91.9596
3	Floyd	Floyd River at James, IA	10750001	06600500	306	2295	42.57666	−96.3114
4	Iowa	Iowa River at Wapello, IA	10580003	05465500	492	32,375	41.17809	−91.1821
5	Little Sioux	Little Sioux River near Turin, IA	10670003	06607500	254	9132	41.96503	−95.9723
6	Maquoketa	Maquoketa River near Green Island, IA	10490005	05418500	249	4022	42.08335	−90.6329
7	Nishnabotna	Nishnabotna River above Hamburg, IA	10360003	06810000	255	7268	40.60167	−95.645
8	Nodaway	Nodaway River at Clarinda, IA	10730001	06817000	274	1974	40.74328	−95.0142
9	Rock	Rock River near Rock Valley, IA	10840001	06483500	276	4123	43.21443	−96.2945
10	Skunk	Skunk River at Augusta, IA	10560002	05474000	262	11,168	40.75365	−91.2771
11	Soldier	Soldier River at Pisgah, IA	10430002	06608500	284	1054	41.83054	−95.9314
12	Thompson	Thompson River at Davis City, IA	10270001	06898000	273	1816	40.64028	−93.8083
13	Turkey	Turkey River at Garber, IA	10220001	05412500	433	4002	42.73999	−91.2618
14	Upper Iowa	Upper Iowa River near Dorchester, IA	10030001	05388250	284	1994	43.42108	−91.5088
15	Wapsipinicon	Wapsipinicon River near De Witt, IA	10820001	05422000	412	6050	41.76697	−90.5349
16	Yellow	Yellow River near Ion, IA	10030002	05389000	364	546	43.11193	−91.2651

, along with other relevant metadata.

A combination of P water quality data and streamflow records was used to estimate stream exports. Because samples have only been collected once per month historically, modeling techniques are needed to predict each river’s P levels on days without sampled data. These daily P concentrations were estimated using Weighted Regressions on Time, Discharge, and Season with Kalman filtering (WRTDSK) at each monitoring site. The widely used WRTDS model identifies relationships between concentration data and several parameters, including streamflow, seasonality, and trends, to estimate long-term loads [52]. The WRTDSK model improves upon the original WRTDS method by applying a Kalman filter, which accounts for serial correlation in model residuals [53]. Sixteen models were successfully simulated using the R EGRET package (https://doi-usgs.github.io/EGRET/) with goodness of fit values ranging from 0.41 to 0.85. Daily loads were calculated using the modeled P concentrations and observed streamflow records and aggregated to annual values. Finally, loads from the 16 sites were summed to calculate Iowa’s P stream export.

2.3. Soil P Storage

The components of the P budget quantify how much P enters and exits Iowa, but they do not reveal how much P is present in Iowa’s soil at any given time. Evaluating the latent P distributed among the fields of Iowa provides a sense of scale compared to the P budget components. This quantity includes all forms of P located within the top layers of the state’s soil. This estimate was not provided in the previous Iowa P budget [29].

The most comprehensive effort to quantify the presence of various chemicals in Iowa’s soil occurred through The Iowa State-Wide Trace Element Soil Sampling Project—published in 2010 as part of the more extensive USGS Geochemical Survey [54]. This project involved collecting soil samples at 532 locations that were uniformly distributed across the state (Figure 2) and analyzing the samples for a number of parameters, including P. The first horizon (shallow) contained the top 0.25 m of soil, while the second horizon (deep) spanned from a depth of 0.25–0.61 m. The concentration of P in the soil was measured in both the shallow and deep horizons at all 532 sites. All soil P concentrations were above the detection limit.

We created histograms of soil P concentrations to explore differences between P levels in the shallow and deep horizons. We also conducted a 2-sample t-test on the datasets from both layers to investigate statistically significant differences (p < 0.05). To estimate the statewide mass of P in the soil, we applied a Kriging spatial interpolation method to the P soil densities measured in the shallow and deep horizons (Figure 2). This produced high-resolution rasters of soil P concentration across Iowa. The amount of P within the horizons was calculated by multiplying the volume of each raster cell by its P concentration and the soil density, which we assumed to be a constant value of 1400 kg/m³ [55]. This calculation yielded statewide estimates of P mass in the shallow and deep soil layers.

2.4. Statistical Summary and Trends

Annual totals for each budget component were compiled, thereby producing a 25-year time series for every budget component, and results are presented in the Supplemental Materials (Table S1). Descriptive statistics were calculated for each time series, including the arithmetic mean (mean), standard deviation (std), coefficient of variation (cv), and several percentile values (min, 25%, 50%, 75%, and max). A Mann–Kendall trend test was conducted on each budget component to determine if there were significant (p < 0.05) monotonic trends from 1998 to 2022 present within the budget. A correlation matrix, which calculated the Pearson correlation coefficients between every budget time series, was also constructed to investigate relationships between individual components (Figure S1).

3. Results

Annual values for each P budget component (in millions of kg) were calculated from 1998 to 2022 (Table S1), and budget statistics are summarized in

Table 2. P budget descriptive statistics (millions kg). Statistics describing the 25-year time series of each budget component. The “trend” row reflects the results of the Mann–Kendall test (‘no trend’ if p > 0.05).

Statistic	Fertilizer	Manure	Human	Industry	Harvest	Grazing	Streams
mean	129.1	106.2	3.3	0.6	226.6	14.7	18.8
std	17.44	12.93	0.11	0.02	24.50	3.66	11.13
min	101	87.3	3.11	0.563	187	9.17	5.26
25%	114	94.9	3.22	0.581	209	11.8	7.25
50%	124	107	3.32	0.6	230	14.3	18
75%	140	119	3.41	0.616	248	17	25.6
max	159	128	3.48	0.629	268	23.4	41.6
cv	0.14	0.12	0.03	0.03	0.11	0.25	0.59
trend	no trend	increasing	increasing	increasing	increasing	decreasing	no trend

.

3.1. Individual Components

The annual values for the individual budget components are plotted in Figure 3. Annual amounts of P usage in fertilizer application ranged from 101 to 159 million kg. P fertilizer loads tend to be dynamic, changing appreciably from year to year, but there was no statistically significant trend in P fertilizer applications. P typically comprised 4–5% of the total fertilizer applied in Iowa.

The P generated from animal manure was similar to the amount purchased in commercial fertilizers, with annual values ranging between 87 and 128 million kg. Swine and cattle are the primary producers of manure P in Iowa, with recent annual values near 70 million kg and 30 million kg, respectively (Figure 4, left). Manure from chicken, sheep, and turkeys was far less (<5 million kg per year). P from manure has significantly increased over the past 20 years (p < 0.05) and can be primarily attributed to the growing hog industry in Iowa.

Based on Iowa’s human population (~3.2 million) and the industrial P factor, annual P loads from industrial sources are near 0.6 million kg. P loads from human waste total slightly above 3 million kg. Both human and industrial components are significantly increasing due to Iowa’s population growth.

The harvested P was the budget’s largest output, ranging from 187 to 268 million kg. Corn (as grain) and soybeans accounted for 93% of the harvested P (Figure 4, right). Annual P output from corn has increased since 1998, with the Iowa corn harvest producing 140 million kg of P per year—a value larger than any other budget component. Soybean values were more constant, with P outputs near 70 million kg. Hay was the next-largest contributor of harvested P, with values around 15 million kg. In contrast to corn, harvested P from hay has declined over the study period. These trends are the result of the widespread transition of pastureland to traditional row crops seen throughout Iowa [56]. There was a statistically significant positive trend for the combined harvested P. The increase in corn production outweighs the decrease in pastureland in terms of P mass, and this effect contributed to an increase in the overall amount of P being harvested over the past 20 years.

Output of P from grazing ranged from 9 to 23 million kg. Grazing P noticeably declined over the budget period, and this was directly attributable to pastureland loss. P harvested from pasture declined from 20 million kg in 2000 to 12 million kg in 2022.

Annual stream export of P ranged from 5 to 42 million kg, with an average load of 19 million kg. The stream export component was highly variable, as river loads fluctuated considerably. The coefficient of variation for river loads (0.59) was far greater than that of the other components, but no statistically significant trend was present. Stream export was correlated with water discharge volume, as more water leaving the state results in greater P export. Hence, river flows and riverine P flux are dynamic from year to year.

3.2. Overall Budget and Net P Balance

Among the individual budget components, P output from harvest was the largest component, with annual values exceeding 200 million kg. Fertilizer and manure inputs routinely exceeded 100 million kg. Other budget categories were small relative to harvest, fertilizer, and manure. Annual grazing and riverine components were approximately 15 million and 19 million kg, respectively, whereas P inputs from industrial and human sources were 0.6 and 3 million kg. While stream P loads were low relative to inputs and harvest, they were the most dynamic of the budget components.

Summing the input and output loads indicates that both input and output totals were consistently above 200 million kg of P per year (Figure 5, top). Both inputs and outputs were increasing over the budget’s timeframe. The net balance of P was estimated by subtracting annual outputs from inputs (Figure 5, bottom). In most years, the net P balance results in a P deficit, meaning that more P exits its landscape than enters it. However, a P surplus was evident in a few years near 2010. Overall, the P net balance estimates suggest that P outputs exceed inputs by approximately 11%.

There was no direct relationship among any of the budget components. An increase in one component did not affect other budget components in the same year. Correlations between stream export and the other six components were low, ranging from −0.21 to 0.12 (Figure S1). The lack of temporal correlation suggests that the budget components act independently of each other. Relationships among budget components are likely the result of outside economic or environmental processes and are not influenced by the behavior of other budget factors.

3.3. Soil Storage of P

Among the 532 soil sample locations from across the state, P concentrations were generally higher in the shallow horizon than in the deep (Figure 6). The average P concentration in the shallow horizon was 770 ppm but was 600 ppm in the deep horizon. Both distributions were relatively normal, and the two-sample t-test confirmed a statistically significant difference in means between the horizons (p < 0.01).

The shallow horizon contained approximately 32 billion kg of P, whereas the deep soil horizon contained nearly 49.5 billion kg of P. Together, approximately 81.5 billion kg of P is contained within the upper 0.61 m of Iowa’s soil. This value includes all types of P, whether bioavailable to plant life or not. With an overall total of 81.5 billion kg, the mass of P in Iowa’s first 0.61 m of soil was almost 300 times greater than the budget’s largest annual loads.

4. Discussion

The annual budget framework greatly expands our understanding of the year-to-year variations that occur with several P components. Other budget efforts have aggregated masses over a 5- to 10-year period [27,29,57], but while this approach captures the broad nutrient behavior spanning those years, it does not assess trends or annual variability. The annual P budget presented herein demonstrates that budget components are not stationary and are impacted by Iowa’s shifting economic activities and demographics. The fluctuations exhibited by several components (e.g., fertilizer application, harvesting, and stream export) suggest that budgets compiled over short timeframes may not be representative of long-term nutrient patterns. Throughout the past 25 years, Iowa has seen population growth, expansion of livestock production, and decreases in pastureland acreage—all impacting the state’s net P balance. Their accumulated effect has changed Iowa’s relationship to P and altered the relative magnitudes of several budget components.

4.1. Iowa’s Net P Balance

Based on the quantification of annual P budget inputs and outputs, Iowa appears to be running a P deficit (negative net annual values) (see Figure 5). However, interpreting this result requires additional context in terms of inputs, outputs, and soil P storage. Overall, the combined annual inputs and outputs across our study’s 25-year window were 239 million and 260 million kg/year, respectively, with outputs exceeding inputs by only 9% over the entire record. For 6 of the 25 years, the net balance was at or above 0, indicating that the annual balance was not uniformly negative. Hence, over the long term (2.5 decades), Iowa’s net P budget could be considered roughly balanced.

Furthermore, the various budget components interact very differently with soil and water. Only the fertilizer, manure, harvest, and grazing components directly interact with Iowa’s agricultural soils. When a net budget is applied to soil-based components ([Fertilizer + Manure] − [Harvest + Grazing]), the inputs and outputs total 235 million and 241 million kg/year, respectively—a difference of only 3%. This suggests an overall balance in the annual P budget associated with Iowa’s agricultural soils.

Although P budget studies often show deficits [29,58,59], our study is unique in providing context for soil P depletion relative to soil P reserves at a regional, statewide scale. Based on systematic soil sampling across the state, Iowa contains approximately 81.5 billion kg of P within the upper 0.61 m of the soil profile. For context, this value is more than 300 times larger than the combined annual P outputs, meaning it would take over 300 years to fully remove the P already present in Iowa’s soils (given Iowa’s current rates of P flux) even if no additional fertilizer or manure inputs were applied. Consequently, the P in Iowa’s top 0.61 m of soil cannot realistically be depleted in the next several decades or beyond, and fluctuations in Iowa’s net P balance have a negligible influence on soil P reserves.

However, it is important to note that much of Iowa’s soil P is not necessarily bioavailable to agricultural commodities [6], so additions of fertilizer and manure P sources are necessary for maintaining crop yields. It may be more appropriate to view fertilizer and manure inputs of P as drivers of economic activity rather than as a means to renew geologic materials. It is unknown how much soil P could be the result of banking excess nutrient application beyond crop needs in previous decades. Termed “legacy” P [60,61], this soil P may be available for crops for many decades in the future.

4.2. Implications for the INRS

The budget provides additional considerations regarding Iowa’s efforts to reduce waterborne P. The foremost of these, the Iowa Nutrient Reduction Strategy (INRS), aims to reduce stream P export by 45% through the implementation of several nutrient-related practices, including fertilizer application guidelines, manure management protocols, augmented point source treatment, and increased erosion control [62]. Stream P loads are influenced by soil erosion [63], which has substantial contributions from on-field and streambank sediment [9].

Stream loads vary tremendously from year to year but are not correlated with the other budget components. Instead, the fluctuations are primarily related to the annual weather and streamflow volume [13]. Increased rainfall triggers erosion, which causes more soil (and more P) to be mobilized and transported to Iowa’s streams. The variability of the rainfall, rather than the magnitude of budget components, impacts yearly stream P.

Given the link between rainfall and stream P loads and the dominant presence of P in Iowa’s soil, meeting the INRS’s P reduction goals will be a matter of successful erosion control. Certain budget components (e.g., industrial and human sources) are minor compared to the total soil P and can only be reduced by a finite amount (~4 million kg/year). Sustainably managing fertilizer application and manure disposal have critical environmental and ecological benefits [64,65], but even eradicating these budget components may not lead to near-term reductions of stream P. There is essentially an unlimited amount of P that can eroded from Iowa’s landscape, and stemming this erosion is imperative in meeting INRS goals. Iowa’s soil P is vast enough that INRS goals will fail if all other P remediation objectives are implemented, but soil erosion is not reduced.

Future research could explicitly investigate the impact of various management scenarios and document their resulting impact on P transport in Iowa [66]. The statewide P budget may be a valuable tool in documenting the cascading impacts of changes to climate, agricultural production, and remediation strategies. These methods could also be applied at a finer resolution, such as at the county or watershed scale [67].

4.3. Budget Limitations

The methods used to assemble this budget work well on large, statewide scales but have limitations when applied to finer spatial resolutions. At the statewide level, local errors largely balance out. Budget components are estimates, but they enable us to confidently gauge each component’s order of magnitude and provide a reasonable picture of P transport since 1998. The calculations for the fertilizer, manure, harvest, and livestock grazing components are easily replicated. Stream P loads are traditionally the most uncertain due to limited water quality data [29], but our methods use the most accurate modeling techniques currently available, which have been demonstrated to work well with datasets spanning multiple decades [68,69].

Several issues arise at a county or watershed scale, and improved datasets would be necessary to scale our methods to such a resolution. Fertilizer distribution is only tracked on a statewide basis. Datasets do not currently contain information on the specific location where each mix is applied. Also, manure is often transported several km before it is disposed of or spread across a landscape [70]. Our analysis assumes manure remains in the same location where it is produced—a strong assumption at the statewide scale but prone to errors evaluating counties or metropolitan areas. Estimates of stream P rely on locating watersheds that align with the budget’s area of interest. Identifying hydrologic regions that describe the majority of Iowa is straightforward, but this may not be the case for smaller study areas that bisect large watersheds.

Another assumption of the budget was that the multiplicative factors used to calculate budget components have remained constant from 1998 to 2022. These factors include the per capita P corresponding to industrial and human sources and the P content of crops and livestock manure. Improvements in wastewater treatment have occurred in Iowa since the initial release of per capita P guidelines [71], suggesting our budget marginally overestimates industrial and human P. Biological engineering may have altered the standard P content with the state’s crops [72], and changes to livestock feed additives may have done likewise to livestock manure [73].

Future budgets may want to reexamine these factors pending their refinement through continued agricultural research. Additional studies could better quantify the range of values associated with the multiplicative factors. While these factors have been noted to be normally distributed [74], the exact parameters of their distributions remain poorly quantified in Iowa [75]. An improved understanding of these factors will help quantify budget uncertainties [76]. While these refinements are unlikely to affect the results of statewide budgets, they become increasingly important at the local scale and will lead to better characterization of P transport within individual watersheds [77].

While this study’s temporal aspect was valuable in understanding the budget components, existing data limitations prevent estimating soil P across various timeframes. Routine soil test P measurements are not widely available, so historical changes to Iowa’s soil P storage cannot be investigated. Therefore, the soil P estimate is one value in our analysis that is assumed to be static.

5. Conclusions

In this study, we calculated each annual component of Iowa’s P budget from 1998 to 2022, a statewide mass balance of all P pathways entering and exiting Iowa’s landscape. Its components consist of four inputs (fertilizer application, manure production, industrial sources, and human sources) and three outputs (harvesting, livestock grazing, and stream export). Our main findings highlighted the dominance of fertilizer application, livestock manure production, and crop harvesting among the budget components. Iowa had a P deficit in most years (20 out of 25). Some temporal trends were present, reflecting the changing agricultural practices in Iowa. P input from manure increased steadily, while P output from livestock grazing saw a constant decline. We also calculated the total amount of P present within the first 0.61 m of Iowa’s soil. The total P mass in this 0.61 m of soil was 81.5 billion kg—over 100 times the size of the largest annual P budget component. This magnitude of this quantity compared to other budget components suggests that P in Iowa’s soil will not be depleted in the near future. Our findings have implications for the INRS. Inhibiting soil erosion will be essential in meeting the state’s nutrient reduction goals due to the vast supply of sediment-bound P in Iowa’s landscape.