Edited by: Paulo Sergio Pavinato, University of São Paulo, Brazil
Reviewed by: Cledimar Rogério Lourenzi, Federal University of Santa Catarina, Brazil; Donnacha Doody, Agri-Food and Biosciences Institute (AFBI), United Kingdom; Carlos Antonio Costa Do Nascimento, São Paulo State University, Brazil
This article was submitted to Soil Processes, a section of the journal Frontiers in Environmental Science
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
The build-up of soil phosphorus (P) beyond plant requirements can lead to a long-term legacy of P losses that could impair surface water quality. Using a database of ∼4,50,000 samples collected from 2001–2015 we report the level of soil P enrichment by soil type, land use and region and the time it would take for Olsen P to decline to agronomic targets (20–0 mg L–1) if P fertilizer was stopped. We also modeled the time it would take for water extractable P (WEP), an indicator of P losses in surface runoff, to decline to an environmental target (0.02 mg L–1). Some 63% of the samples were enriched beyond agronomic targets. The area-weighted median time to reach the agronomic target was predicted to occur within a year for 75% of samples but varied up to 11.8 years in some land uses. However, the area-weighted time to reach an environmental target was 26–55 years for the 50th and 75th percentile of areas. This indicates that while an agronomic target can be easily met, additional strategies other than stopping P fertilizer inputs are required to meet an environmental target.
The loss of phosphorus (P) from land can impair surface water quality (
The addition of fertilizer-P to soils aims to establish optimal STP concentrations. Due to differences in ASC or other measures of soil P sorption, recommended fertilizer rates to reach agronomic optima vary by soil and crop type. However, it has been common practice to apply a little extra P when cash is available as an insurance policy against poor years when little P may be applied (
The most extreme version of a negative P balance involves no application or returns of P to the soil. Research has shown that in grassland systems with little erosion and a similar climate this no-P scenario results in a rate of decline that reflects the initial STP concentration and ASC (
The drive for profit is a common factor in rapid land use change. In New Zealand, the number of dairy cows increased from 2.9 M in 1996 to 5.0 M in 2016, much of this has been onto land traditionally used for drystock sheep and beef grazing (
The objective of our paper was to determine: (1) the state as of 2016 and any trends over the last 15 years in soil Olsen P concentrations, and (2) to determine the time it would take soils enriched with P to decrease to an agronomic or environmental optimum for their land use and soil group.
Data were obtained for soil Olsen P concentrations from the three major commercial laboratories in New Zealand (ARL, Eurofins, Hills Labs). This data comprises soil test results for various commercial agricultural enterprises undertaken to assess soil fertility and fertilizer requirements. The database covered the years 2001–2015 and contained ∼4,50,000 data points. Each laboratory has international standards (ISO 9000) relevant to the collection, analysis and recording of Olsen P data. All dairy and drystock samples were from the 0–7.5 cm depth, while cropping samples were from the 0–15 cm depth.
Each sample had metadata describing region (16 local authorities) and land use [dairy, drystock (sheep, sheep and beef, and red deer), cropping or horticulture]. The Eurofins and Hills Labs data (c. 2,30,000 samples) had data for soil group (ash, peat, pumice, and sedimentary). These groups corresponded to New Zealand soil orders (
We estimated the water extractable P (WEP) concentration (mg L–1) of each sample using the equation of
As input data we combined median region-by-land use-by-soil group-specific Olsen P concentrations with location specific ASC data mapped nationally at a 1:50,000 scale within the New Zealand fundamental soil layer (
We used targets for soil Olsen P concentrations that corresponded to an agronomically optimum level of production and a WEP concentration set as an environmental target unlikely to induce excessive levels of periphyton biomass in streams and rivers.
Agronomic targets were set for different land uses within each soil group (
Median concentrations of soil Olsen P and WEP were derived for the 2010–2015 years. This range was chosen as an indicator, unlikely to be influenced by trends, of the present state of these parameters in each region by land use by soil group combination. Because the data were skewed, concentrations for each region, soil group and land use, and their interactions, were compared as medians using an analysis of variance of ranked data. Data from 2001–2015 for each region by land use by soil group combination with >500 data values were analyzed for trends in the annual enrichment or depletion of median Olsen P concentrations using a non-parametric Mann-Kendall test within the Time Trends software package (
To estimate the time to reach an agronomic and environmental target we used the equations of
We calculated the time (
where WEP
We calculated the time (
Equations 1–3 were calibrated by
Data for the time to deplete the median region by soil group by land use combination for soil Olsen P or WEP to their respective agronomic or environmental target in each combination or at the land use, soil group or national level were also calculated as area-weighted medians. Areas for each combination are given in
Within the database, Peat, Pumice, Sedimentary and Volcanic soils had 7,561, 16,109, 3,42,537, and 54,749 samples in them, respectively. Most soils were sampled under dairy land (2,21,035), followed by drystock (1,41,775), cropping (47,398) and horticulture (10,748). The median number of samples within a region by land use by soil group combination was 380 while the mean was 3,028. The number of samples measured per year was slightly greater in later years.
Median Olsen P concentrations across region by land use by soil group combinations are given in the
Median Olsen P concentrations (mg L–1) in each region by soil group by land use (Hort = horticulture) combination.
Auckland | 31 | 52 | 34 | – | – | – | – | – | 40 | 38 | 22 | 26 | 30 | 40 | 19 | 9 |
Bay of Plenty | 23 | 43 | 29 | 43 | 47 | 47 | 25 | 48 | 36 | 39 | 18 | 49 | 48 | 41 | 22 | 40 |
Canterbury | –1 | 28 | 13 | – | – | – | – | – | 22 | 28 | 21 | 24 | – | – | – | – |
Gisborne | – | – | – | – | – | 19 | – | – | 34 | 25 | 19 | 41 | – | – | – | – |
Hawkes Bay | – | – | – | – | 19 | 22 | 17 | 34 | 25 | 29 | 20 | 36 | 13 | 32 | 20 | 29 |
Manawatu/Wanganui | – | 24 | – | – | – | 22 | 7 | – | 25 | 31 | 19 | 94 | 20 | 38 | 18 | 22 |
Marlborough | – | – | – | – | – | – | – | – | 25 | 28 | 20 | 22 | – | – | – | – |
Nelson | – | – | – | – | – | – | – | – | – | 29 | 23 | – | – | – | – | – |
Northland | 33 | 42 | 6 | – | – | – | – | – | 29 | 37 | 24 | 38 | 32 | 32 | 34 | 34 |
Otago | – | 26 | 9 | – | – | – | – | – | 21 | 28 | 19 | 21 | – | – | – | – |
Southland | 20 | 18 | 17 | – | – | – | – | – | 20 | 27 | 21 | 24 | – | – | – | – |
Taranaki | – | 38 | – | – | – | – | – | – | 29 | 32 | 20 | – | 24 | 41 | 26 | 106 |
Tasman | – | – | – | – | – | – | – | – | 24 | 29 | 21 | 52 | – | – | – | – |
Waikato | 33 | 43 | 23 | 21 | 44 | 48 | 25 | 33 | 31 | 43 | 21 | 34 | 26 | 40 | 20 | 29 |
Wellington | – | – | – | – | – | – | – | – | 25 | 32 | 19 | 23 | – | – | – | – |
West Coast | – | – | – | – | – | – | – | – | 11 | 29 | 20 | – | – | – | – | – |
Of the 124 combinations, 32 showed significant trends in concentrations from 2001 to 2015 (
Mean annual percentage change in the median concentration of Olsen P for each region by soil group by land use (Hort = horticulture) combination over 2001–2015.
Auckland | ns1 | 2.5% | ns | – | – | – | – | – | 6.6% | −0.9% | ns | ns | ns | −0.9% | ns | ns |
Bay of Plenty | ns | ns | ns | ns | ns | ns | −0.6% | ns | ns | ns | −5.2% | ns | ns | −1.8% | ns | 1.4% |
Canterbury | –2 | ns | ns | – | – | – | – | – | ns | ns | ns | ns | – | – | – | – |
Gisborne | – | – | – | – | ns | – | – | ns | ns | ns | ns | – | – | – | – | |
Hawkes Bay | – | – | – | ns | ns | −4.0% | ns | 5.3% | ns | 0.1% | 3.4% | ns | ns | ns | ns | |
Manawatu/Wanganui | – | ns | – | – | – | ns | ns | – | 2.1% | ns | 3.0% | ns | ns | 1.3% | –2.1% | 16.2% |
Marlborough | – | – | – | – | – | – | – | – | ns | −3.2% | 1.3% | ns | – | – | – | – |
Nelson | – | – | – | – | – | – | – | – | ns | ns | ns | – | – | – | – | – |
Northland | ns | 1.3% | ns | – | – | – | – | – | ns | −0.5% | 1.7% | 4.2% | ns | −1.6% | 5.3% | ns |
Otago | ns | ns | – | – | – | – | – | ns | −1.1% | ns | ns | – | – | – | – | |
Southland | ns | ns | ns | – | – | – | – | – | 1.3% | ns | ns | ns | – | – | – | – |
Taranaki | – | ns | – | – | – | – | – | – | ns | ns | ns | – | ns | ns | ns | ns |
Tasman | – | – | – | – | – | – | – | – | ns | ns | ns | 5.2% | – | – | – | – |
Waikato | ns | ns | ns | ns | ns | 0.8% | ns | ns | ns | ns | −1.6% | ns | −1.8% | −0.3% | ns | ns |
Wellington | – | – | – | – | – | – | – | – | ns | ns | ns | ns | – | – | – | – |
West Coast | – | – | – | – | – | – | – | – | ns | ns | ns | – | – | – | – | – |
The percentage of samples collected in 1988–1996, 1996–2001, and 2002–2015 exceeding target ranges in Olsen P for each soil group in dairy or sheep and beef land use. Data for 1988–2001 and targets (45, 45, 25, and 30 mg L–1 for Peat, Pumice, Sedimentary, and Volcanic soils, respectively) were taken from
Median Olsen P concentrations were above their respective soil group-by-land use-specific agronomic target in 82 of the 124 combinations (
Median and the 75th percentile (in parentheses) of the time (years) taken for Olsen P to decrease to agronomic targets in each region by soil group by land use (Crop = Cropping, Hort = horticulture) combination if fertilizer P was withheld.
Auckland | 2.4 (2.6) | 1.7 (2.6) | 0 (0) | – | – | – | – | – | 6.3 (8.1) | 3.3 (3.3) | 0.6 (1.0) | 3.6 (4.4) | 6.1 (6.1) | 4.9 (4.9) | 0 (0) | 0 (0) |
Bay of Plenty | 0 (0)1 | 0.5 (0.6) | 0 (0) | 4.1 (5.3) | 4.5 (5.5) | 1.2 (1.2) | 0 (0) | 4.6 (4.6) | 5.8 (6.4) | 3.6 (3.9) | 0 (0) | 7.0 (7.0) | 6.9 (9.0) | 4.6 (4.9) | 0 (0) | 8.1 (8.1) |
Canterbury | –2 | 0 (0) | 0 (0) | – | – | – | – | – | 1.4 (1.4) | 0.7 (0.7) | 0.3 (0.5) | 1.8 (1.8) | – | – | – | – |
Gisborne | – | – | – | – | – | 0 (0) | – | – | 3.9 (5.5) | 0 (0) | 0 (0) | 4.4 (6.4) | – | – | – | – |
Hawkes Bay | – | – | – | – | 0 (0) | 0 (0) | 0 (0) | 2.3 (2.7) | 2.1 (2.1) | 0.9 (1.4) | 0 (0) | 3.4 (3.4) | – | 3 (3) | 0 (0) | – |
Manawatu/Wanganui | – | 0 (0) | – | – | – | 0 (0) | 0 (0) | – | 2.5 (3.3) | 1.2 (1.9) | 0 (0) | 3.7 (7.2) | 1.4 (1.4) | 4.4 (4.9) | 0 (0) | 2.8 (2.9) |
Marlborough | – | – | – | – | – | – | – | – | 2.1 (2.1) | 1.1 (1.1) | 0 (0) | 1.4 (1.4) | – | – | – | – |
Nelson | – | – | – | – | – | – | – | – | – | 1.4 (1.5) | 1.5 (1.6) | 0 (0) | – | – | – | – |
Northland | 0.4 (2.4) | 0.2 (0.3) | 0 (0) | – | – | – | – | – | 2.7 (5.3) | 1.9 (3.2) | 1.2 (1.9) | 6.1 (7.7) | 6.3 (6.3) | 2.9 (3.2) | 4.7 (5.3) | 7.0 (7.8) |
Otago | – | 0 (0) | 0 (0) | – | – | – | – | – | 0.8 (1.1) | 0.6 (0.7) | 0 | 0.4 (0.8) | – | – | – | – |
Southland | 0 (0) | 0 (0) | 0 (0) | – | – | – | – | – | 1.4 (1.4) | 0.5 (0.7) | 0.3 (0.4) | 3 (3) | – | – | – | – |
Taranaki | – | 0 (0) | – | – | – | – | – | – | 5.2 (5.5) | 2.6 (2.9) | 0 (0) | 0 (0) | 3.8 (3.9) | 5.4 (5.5) | 2.2 (2.2) | 11.5 (11.5) |
Tasman | – | – | – | – | – | – | – | – | 1.9 (3.2) | 1.4 (1.4) | 0.5 (0.5) | 3.9 (4.8) | – | – | – | – |
Waikato | 2.7 (3.4) | 0.6 (0.8) | 0 (0) | 0 (0) | 4.2 (4.2) | 1.3 (1.3) | 0 (0) | 2.5 (2.5) | 4.8 (6.0) | 4.1 (5.3) | 0.5 (0.6) | 5.4 (6.8) | 4.6 (4.6) | 5.2 (5.2) | 0 (0) | 6.1 (6.1) |
Wellington | – | – | – | – | – | – | – | – | 2.1 (3.1) | 1.3 (2.2) | 0 (0) | 1.7 (2.6) | – | – | – | – |
West Coast | – | – | – | – | – | – | – | – | 0 (0) | 1.1 (1.2) | 0 (0) | – | – | – | – | – |
Maps showing the distribution of median topsoil Olsen P (mg L–1) concentrations
Across all soil groups, the longest area-weighted time to reach the agronomic target was in the Volcanic soils (3.3 years), followed by the Pumice (2.3 years), Peat and Sedimentary (<1 year) soils. Amongst land uses, horticultural soils were estimated to take the longest time to reach the target (3.3 years) followed by dairy, cropping and drystock soils at 2.3, 1.9, and <1 year, respectively. The area-weighted time for the 75th percentile was 1 year and the corresponding time to reach the Olsen P target was 3.7, 2.6, 1.2, and 1.2 years for Volcanic, Pumice, Sedimentary and Peat soils, respectively.
Compared to the time to reach an agronomic target, the median time to reach the environmental target was much greater: the national area-weighted time was 26 years and ranged from 0 years for land uses of Peat soils in Southland to >100 years for six land uses of Volcanic soils (
Median and the 75th percentile (in parentheses) of the time (years) taken for water extractable P (WEP) to decrease to environmental targets in each region by soil group by land use (Crop = Cropping, Hort = horticulture) combination if fertilizer P was withheld.
Auckland | 48 (100) | 23 (34) | 23 (27) | 25 (34) | 26 (26) | 24 (39) | 33 (72) | 64 (64) | 39 (39) | 0 (23) | 0 (0) | |||||
Bay of Plenty | 37 (37) | 25 (29) | 30 (54) | 25 (32) | 24 (28) | 24 (24) | 34 (34) | 24 (24) | 26 (29) | 25 (28) | 25 (52) | 24 (24) | 24 (29) | 30 (33) | 40 (89) | 34 (34) |
Canterbury | 24 (24) | 31 (31) | 23 (23) | 22 (23) | 24 (42) | 23 (23) | ||||||||||
Gisborne | 76 (76) | 23 (27) | 24 (34) | 28 (35) | 22 (25) | |||||||||||
Hawkes Bay | 36 (52) | 39 (100) | 64 (64) | 24 (27) | 23 (23) | 22 (30) | 24 (33) | 22 (22) | 88 (88) | 100 (100) | ||||||
Manawatu/Wanganui | 23 (36) | 40 (40) | 77 (77) | 24 (34) | 22 (29) | 25 (35) | 22 (23) | 100 (100) | 36 (52) | 0 (55) | 100 (100) | |||||
Marlborough | 23 (23) | 31 (31) | 24 (45) | 23 (23) | ||||||||||||
Nelson | 30 (33) | 37 (49) | 47 (72) | |||||||||||||
Northland | 28 (28) | 22 (25) | 18 (74) | 22 (44) | 22 (26) | 23 (35) | 26 (36) | 46 (46) | 44 (82) | 41 (69) | 40 (67) | |||||
Otago | 23 (23) | 18 (22) | 23 (24) | 22 (24) | 24 (34) | 21 (23) | ||||||||||
Southland | 0 (24) | 0 (21) | 0 (21) | 24 (44) | 23 (32) | 24 (43) | 35 (35) | |||||||||
Taranaki | 40 (52) | 44 (59) | 38 (55) | 27 (100) | 0 (18) | 100 (100) | 42 (46) | 100 (100) | 23 (23) | |||||||
Tasman | 23 (45) | 30 (30) | 42 (42) | 21 (22) | ||||||||||||
Waikato | 32 (70) | 24 (35) | 23 (37) | 24 (24) | 24 (24) | 24 (24) | 34 (34) | 28 (28) | 29 (49) | 25 (32) | 24 (42) | 27 (42) | 100 (100) | 49 (49) | 0 (45) | 100 (100) |
Wellington | 23 (28) | 22 (28) | 25 (35) | 23 (37) | ||||||||||||
West Coast | 43 (66) | 26 (30) | 27 (45) |
Some care is needed when interpreting our dataset. There are questions over bias, representativeness, and the repeatability of measurements. Although we have isolated soil Olsen P concentrations to the post code level, no information was available to determine if the samples submitted represented the range of Olsen P concentrations present within a post code or within a land use by soil group combination. We also had no information to determine if a disproportionate number of samples were being sent from enterprises with low or high Olsen P concentrations or if multiple samples were coming from the same enterprises over time. However, analysis of data from one of the providers, ARL, enabled the enterprise to be identified in about 20,000 samples. Of these samples, <5% of enterprises were represented more than once on a sampling date or had multiple samples over the 15-year period of record.
Bias is also possible due to errors in recording a sample’s location and assignment to the appropriate soil group. Related to this, the calculation of timeframes to deplete soil P to agronomic and environmental targets is subject to errors associated with the use of default ASC values for specific soil types. The original equations generated by
The majority of region by soil by land use by soil group combinations had median concentrations that were close to their agronomic target. However, of those soils with an Olsen P concentration above the agronomic target a large proportion were in Volcanic dairy soils (
There were few trends in Olsen P concentrations. This is despite significant land use change in some regions. For instance, dairy cattle numbers in Canterbury, Otago, and Southland increased from 5,76,632 to 1,767,763 from 2001 to 2015 (
Following the cessation of fertilizer-P additions the rate of decline in Olsen P has been shown to vary according to soil type and the initial STP concentration (
Much like Olsen P, the rate of decline in WEP is a function of the initial STP concentration of the soil and soil P sorption. For example, in addition to predicting the rate of decline in WEP was a function of Olsen P and ASC,
On those soils where Olsen P is above the agronomic targets, Olsen P concentrations must reduce. The first step in making reductions is to ensure that the on-farm P balance is not in surplus thereby avoiding soil P enrichment. While livestock systems with confinement will likely have P a surplus (
While a P balance is a good method of decreasing Olsen P toward a target, it can take many years to do so. There are methods that can achieve reductions in Olsen P faster. For instance, plowing to reduce the soil P stratification can decrease topsoil P concentrations by half – overnight, abating surface runoff losses (
In pasture systems,
Grain crops and grass forage rotations were able to deplete topsoil STP by 11 – 36% over 7–16 years in Swedish lysimeter trials. However, translocation of P to the subsoil was noted in three of four soil types with associated increases in P leaching (
We predicted that the area-weighted time for median and 75th percentile of Olsen P concentration across New Zealand to decrease to its agronomic target was within a year. However, the equivalent area-weighted time for WEP was 26–55 years, inferring that to reach targets for WEP would require the cessation of P fertilizers over generational timeframes. Pursuing low WEP concentrations may also result in Olsen P concentrations lower than the agronomic target in many soils, impairing production. For instance, in a sheep grazed trial in Winchmore, Canterbury, New Zealand withholding fertilizer for 6 years resulted in a 34–53% decrease in pasture production (
Other strategies are therefore required to augment or replace the cessation of P fertilizer to soils to reduce WEP concentrations. Such strategies include, but are not limited to: the application of Al-, Fe- or Ca-P-sorbing compounds like alum, bauxite or crushed shells to reduce the dissolution of P into soil solution (
Analysis of the P status of New Zealand agricultural soils revealed that 63% of the productive land area had Olsen P concentrations more than the agronomic optimum. Accumulation of soil P was found to be related to both the soil order and farming system with dairy operations on volcanic soils, with high ASC values, exhibiting the greatest exceedances. If fertilizer application was halted, we predicted that the median Olsen P concentration could be reduced to the agronomic optimum within 1 year. In contrast, reducing the median WEP to an environmental target would require 36–55 years for most soil orders and farming systems, but greater than 100 years for volcanic soils with the greatest P enrichment. To safeguard water quality, strategies such as phytomining are required to accelerate the decline in WEP from the most enriched soils along with mitigation strategies to reduce P solubility. The discrepancy between the timescales to reach these two targets highlights the difficulty in maintaining high agronomic productivity while ensuring good water quality status. Revision of the agronomic targets should be undertaken to account for increases in P use efficiency in modern farming systems and could go some way to closing this gap. However, our results suggest that new land management strategies will be required to develop agronomically and environmentally sustainable farming systems which are productive at a lower Olsen P status.
The datasets generated for this study are available on request to the corresponding author.
RM wrote the manuscript and analyzed the data along with AN. RD contributed text to the manuscript. PP developed the maps and generated spatial data.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This work was supported by the Our Land and Water National Science Challenge (contract C10×1507 from the Ministry of Business, Innovation and Employment).
We are grateful to Hills Laboratories, ARL, and Eurofins for the supply of Olsen P data.
The Supplementary Material for this article can be found online at: