Executive Summary
Phosphogypsum (PG), a by-product of phosphoric acid manufacture, is primarily gypsum (CaS04.2H20). Mined gypsum as well as PG have been established to be good sources of S and Ca for crops (Shainberg et al., 1989; Alcordo and Rechcigl, 1993; and Alcordo and Rechcigl, 1995). Unlike mined gypsum, PG may contain As, Ba, Cd, Cr, Pb, Hg, Se, Ag, and F, any one of which can be toxic at high concentration levels. Phosphogypsum is also highly acidic (pH>2 to <5) compared with mined gypsum (pH=”7),” and the high solubility of PG (2.6 g L-1) relative to that of calcium carbonate (0.015 g L-1) may affect the quality of surficial groundwater in terms of electrical conductivity (Ec) or total dissolved solids (TDS). However, it is primarily because of the presence of small amounts of radionuclides in PG, particularly radium-226 (226Ra) which is the source of the gas radon-222 (222Rn), that the United States Environmental Protection Agency (USEPA) has imposed severe restrictions on PG use.
Four field experiments were conducted using forage grasses. Two experiments (one agronomic-environmental and one solely agronomic) were conducted using Pensacola bahiagrass (Paspalum notatum Flugge) on an established pasture. Two similar experiments were conducted using annual ryegrass (Lolium multiflorum Lam.) on an annually seeded field. All four experiments were conducted to determine the effect of PG, applied as a source of S and Ca, on forage yields and quality. Nonradiological environmental and radiological data were collected from the agronomic-environmental experiments to evaluate the environmental impacts associated with PG use in agriculture on soil, groundwater, plant tissue, and on the emanation of 222Rn to the atmosphere. The studies were conducted by the University of Florida (UF), Institute of Food and Agricultural Sciences (IFAS) at the Range Cattle Research and Education Center (RCREC) at Ona, Florida. The soils used were Florida Spodosols (Myakka series in the case of the bahiagrass field and Pomona series in the case of the annual ryegrass field). The experimental rates of O.O, 0.4, 2.0, and 4.0 Mg (Mg = megagram = 106 grams) PG ha-1 were broadcast by hand over the experimental plots. In the bahiagrass experiments, the PG was not mixed with the soil but was left on the surface and allowed to dissolve and leach naturally down the soil profile by rain. In the ryegrass plots, the PG was mixed with the top 15-cm of the soil using a disk prior to seeding. The 0.4 Mg PG ha-1 treatment was applied annually for 3 years and the 2.0 and 4.0 Mg PG ha-1 treatments were applied only at the beginning of the study.
The experiments ran from 1990 to 1993. The radiological analyses, except for ambient Rn and gamma radiation, were performed by commercial environmental laboratories. The nonradiological analyses were done by various UF-IFAS laboratories. Ambient Rn and gamma radiation, which were measured using electret ion chambers (EIC) were determined at the RCREC lab. All data were statistically analyzed by the Statistics Department of UF-IFAS.
This report (Volume II) covers the agro-environmental and radiological aspects of
the study. The agronomic results are reported in Volume I.
Agro-environmental Aspects. The pH in water of the PG used in the study (1:1) was
4.6. The PG dissolved at a constant rate of 2.6 g L-1 in water and 4.3 g L-1 in a
mixed acid solution of 0.025 M HCl and 0.0125 M H2SO4 (Mehlich 1). It contained small
amounts (mg kg-1) of As (5.0), Ba (46.0), Cd (0.7-1.1), Cr (2.9), Pb (4.0), Hg (<0.01),
Se (<0.05-1.6), and Ag (<0.2-2.0). Fluoride content was about 0.43 %. The levels of
226Ra, 210Pb, and 210Po activities were 18, 31, and 24 pCi g-1, respectively.
The heavy metals (HMs) As, Ba, Cd, Cr, Pb, Hg, Se, and Ag are used by USEPA to determine whether a solid waste is hazardous or not under its “toxicity characteristic” (TC) category based on each metal’s leaching potential (LP) from a solid waste, determined according to USEPA procedure. Using the extensive data of May and Sweeney (1983) on measured LPs or TCs of the HMs in Florida PG samples from 9 PG stacks, the LP or TC for each heavy metal in any Florida PG may be estimated by:
Estimated LP or TC (unit: mg HM L-1 of leachate) = [(measured mg HM kg-1 of PG used in the study)/(measured avg mg HM kg-1 in Florida PGs with measured avg LP or TC)] x [measured avg LP or TC of Florida PGs].
The estimated LPs or TCs of the various HMs in the PG used in the study were 0.12 (As), 0.09 (Ba), 0.02 (Cd), 0.02 (Cr), 0.03 (Pb), <0.001 (Hg), <0.001 (Se), and 0.30 (Ag). Compared with the USEPA’s TC limits of 5.0 (As), 100 (Ba), 1.0 (Cd), 5.0 (Cr), S.0 (Pb), 0.2 (Hg), 1.0 (Se), and 5.0 (Ag) mg L-1, the estimated LP values for any HM in the PG used in the study showed that the PG did not fall under the “toxicity characteristic” category of a USEPA hazardous waste.
The projected increases in heavy metal concentrations at the 0-15 cm surface soil (assumed bulk density = 1.5 g cm-1) in the ryegrass plots due to application of 4.0 Mg PG ha-1 were estimated to be 0.010 (As), 0.092 (Ba), 0.002 (Cd), 0.006 (Cr), 0-008 (Pb), <0.0001 (Hg), 0.003 (Se), and 0.004 (Ag) mg kg-1 soil. Twelve months after PG application, Cd, Cr, Pb, Hg, and Se were determined in the soil profile down to 90 cm depth at 15-cm depth intervals. The measured ranges of concentrations in plots that received 4.0 Mg PG ha-1 for Cd, Cr, Pb, Hg, and Se were 0.04-0.10, 0.40-0.80, 0.14-0.S8, 0.00-0.16, and 0.00-0.16 mg kg-1 soil, respectively. The statistics showed no difference between treatments at any soil depth. These indicated that the HM contaminants in PG used in the study would not present any real short-term or long-term environmental concern for the soil.
Assuming that the LPs of the HMs in PG now mixed with the soil were the same as in pure PG, the TCs of the top 0-15 cm soil were estimated to increase by 0.0002 (As), 0.0002 (Ba), 0.00003 (Cd), 0.00003 (Cr), 0.00005 (Pb), 0.0000003 (Hg), 0.000006 (Se), and 0.0003 (Ag) mg L-1. Relative to the USEPA primary standards for drinking water which are 0.05 (As), 1.00 (Ba), 0.01 (Cd), 0.05 (Cr), 0.05 (Pb), 0.002 (Hg), 0.01 (Se), and 0.05 (Ag) mg L-1 the projected increments are simply insignificant. Hence, it is concluded that the estimated increases in the LPs or TCs due to these heavy metal contaminants in PG would not present any impact, short- or long-term, on the quality of the surficial groundwater.
The statistical analysis of the pH of surficial groundwater from the bahiagrass experiment averaged annually (1990, 1991, 1992) and over the 3-year period (1990-1992) indicated no effect of PG treatments on the pH of surficial groundwater sampled at the surface (runoff) and at 60 and 90 cm depths. The annual average pH of runoff during 3 years of sampling ranged from 4.9 to 5.4 for the control plots and from 4.3 to 6.3 for all treated plots. Groundwater sampled at 60 cm depth had pH ranging from 4.5 to 4.8 for the control plots and from 4.6 to 5.6 for all treated plots. Groundwater sampled at 120 cm depth had pH ranging from 4.9 to 5.2 for the control plots and 4.0 to 5.8 for all treated plots.
The statistical analysis of the pH of groundwater samples averaged annually (1990-91 and 1991-92) and over the 2-year period (1990-91 to 1991-92) at 60 and 120 cm depths from the ryegrass experiment also showed no effect of PG applications. The annual average pH of groundwater from the control plots ranged from 4.0 to 4.4 at 60 cm depth and from 4.3 to 4.9 at 120 cm depth. For all treated plots, groundwater pH ranged from 3.7 to 4.4 and from 4.1 to 4.9 at 60 and 120 cm depths, respectively.
In the bahiagrass experiment, the highest annual (1991) average Ec of 1190 and 3-year
average of 821 umho cm-1 were noted in groundwater sampled at 60 cm depth from plots
that received 4.0 Mg PG ha-1. In the ryegrass experiment, the highest annual (1990-91)
and 2-year average Ec of 1421 and 1137 umho cm-1, respectively, were again noted in
groundwater from plots treated with 4.0 Mg ha-1 sampled at the 60 cm depth. The highest
value of 1421 umho cm-1 observed in the ryegrass experiment, however, was still less
than the upper Ec limit of 1500 umho cm-1 for potable water in the United States.
Converting the highest Ec value of 1421 umho cm-1 into TDS (Ec x 0.66) gave 938 mg
of dissolved solids L-1 which, again, was well below Florida’s TDS standards of <3000
mg L-1 for water for agricultural use and <1000 mg L-1 for domestic and industrial
uses.
In the case of F levels in groundwater from the bahiagrass experiment, the highest annual and 3-year average F levels in surficial groundwater from all depths were 0.83 and 0.36 mg L-1, respectively. Both values were observed in water samples collected at 60 cm depth from plots that received 2.0 Mg PG ha-1. In the tilled ryegrass plots, the highest annual and 2-year average F levels in groundwater from all treated plots and for all depths (60 cm and 120 cm) were 0.53 and 0.43 mg L-1, respectively. The highest annual average F levels in groundwater sampled at 60 cm and at 120 cm depths from the control plots were 0.31 and 0.19 mg L-1, respectively.
It is widely accepted that approximately 1.0 Mg F L-1 in drinking water can effectively reduce dental caries without harmful effects on health. The Florida drinking water primary standards allow for a maximum contamination level (MCL) for F from 1.4 to 2.4 mg L-1. None of the individual measured F values in the bahiagrass and ryegrass experiments exceeded the MCL values. The annual and the 3-year (bahiagrass) or 2-year (ryegrass) average values were even less than 1.0 mg F L-1. Therefore, it is concluded that application of PG up to 4.0 Mg ha-1 to an established pasture or tilled land would not lead to unacceptable levels of F in surficial groundwater.
Radiological Aspects: Established Bahiagrass Pasture.
Phosphogypsum contatining 18, 31, and 24 pCi g-1 of 226Ra, 210Pb and 210Po, respectively, was applied to the experimental plots which also contained measured activities of 0.55, 0.61, and 0.53 pCi g-1 in the top 15-cm layer of untreated soil. It is calculated that each Mg PG ha-1 added 1820, 3080, and 2430 pCi m-2 and increased the activities in the upper 15-cm layer (assumed density 1.5 g cm-3) by 0.008, 0.014, and 0.011 pCi g-1. For the maximum treatment of 4.0 Mg ha-1, the calculated increases in the initial activities of 226Ra, 210Pb, and 210Po, averaged over the 15cm layer, were 0.032, 0.055, and 0.043 pCi g-1, respectively. These small additions to the natural activity in the soil could not be detected. An increase in 210Pb due to the 4.0 Mg PG ha-1 treatment was detected in the top 15 cm layer in the first year of sampling, but it could not be detected in subsequent years. Any downward transport of the three radionuclides over the three years could not be detected.
There was no consistent correlation of gamma-radiation level with PG treatment. Any influence of PG at treatment levels up to 4 Mg ha-1 is insignificant in comparison to temporal and spatial variations in the local gamma-radiation background.
There was limited evidence of measurable increases in 226Ra uptake by bahiagrass due to PG application. However, contributions by PG at the application levels used are small relative to the variation in the background level of 226Ra and were difficult to measure. In regrowth forage, increases in 226Ra uptake by bahiagrass during the first three years were estimated to range from 0.01 to 0.02 pCi g-1 per Mg PG ha-1. In hay forage, the data indicated an uptake in the order of 0.015 to 0.017 pCi g-1 per Mg PG ha-1. No measurable effect of PG on 210Pb uptake by bahiagrass was observed. For 210Po, there was only limited evidence of increased uptake at the PG treatment levels used, and any contribution was small relative to the variations in the background levels. For 210Pb and 210Po, the data were insufficient to support calculation of an uptake factor.
Water sampling included runoff and also groundwater from 60-cm and 120-cm depths. 226Ra appeared in runoff during the first three years after PG application. The data suggested, but did not confirm, an effect at 60 cm and 120 cm. However, the maximum activity observed in water, 1.8 pCi L-1, was well below the current (3 pCi L-1) and proposed (20 pCi L-1) drinking water standards for 226Ra. The data for 210Pb in water were equivocal; there was very limited evidence for an effect of PG treatment on the well water. However, comparison of observed levels (up to 2.5 pCi L-1) to proposed criteria (1 pCi L-1) indicated that the potential for 210Pb in water deserves further evaluation. No effect of PG on 210Po in groundwater could be detected by this experiment for treatments up to 4 Mg ha-1. The maximum values for treated and control plots, 1.5 and 1.8 pCi L-1, respectively, were less than the current 15 pCi L-1 standards for gross alpha activity in drinking water. Radon flux is a sensitive parameter for evaluating Rn potential of land. No significant effects of the PG treatment on Rn flux were detected. Any increase attributable to PG treatment at levels up to 4 Mg ha-1 is insignificant in relation to the variations in the natural levels of Rn flux. Consequently, any effect on indoor Rn in future structures built over lands treated at this level will be insignificant relative to the variations experienced in indoor Rn levels.
Radiological Aspects: Tilled Land cropped to Annual Ryegrass.
As with the bahiagrass experiment, the small additions of 226Ra, 210Pb, and 210Po to the natural radioactivity of the soil could not be consistently detected. No PG-attributable increases in background gamma radiation could be detected.
As in the case with bahiagrass, there was limited evidence of measurable 226Ra uptake by ryegrass forage related to the PG treatment; the uptake factor was estimated to be on the order of 0.039 pCi g-1 per Mg ha-1 for regrowth and mature forage. There was an indication but no statistically-significant evidence, of PG-attributable 210Pb and 210Po uptake for treatment rates up to 4 Mg ha-1.
Groundwater at a 120-cm depth was sampled five times. The data suggested that PG treatment of the soil increased the 226Ra activity in groundwater; however, effects were statistically significant only during the first year after treatment, and the treated plot results did not correlate well with treatment level. The maximum 226Ra activity, 2.35 pCi L-1, was below the current and proposed drinking water standards of 3 pCi L-1 and 20 pCi L-1, respectively. The treated plots tended to have higher activities of Z10Pb than the controls; however, after the first year there was no consistent, meaningful correlation of activity with treatment level. Comparison of the observed levels (up to 2.5 pCi L-1) to proposed criteria (1 pCi L-1) indicated that the potential for 210Pb in groundwater deserves further attention. For 210Po in ground water, no significant effects of PG treatment up to 4 Mg ha-1 were observed, and all activities were well below the current and proposed 15 pCi L-1 standards for gross alpha activity in drinking water.
Soil surface Rn flux was determined six times at the test plots. PG treatment increased Rn flux; the estimated contribution is on the order of 0.002 pCi m-2 s-1 per Mg PG ha-1.
Ambient atmospheric Rn activities 1 m above the soil surface were determined 13 times after treatment application for periods ranging from 30 to 90 days. No effects of PG treatment were observed; however, Rn activity 1 m above an individual small plot is probably not representative of the Rn emission from that plot because of atmospheric mixing.
Overall, the radiological impact on the environment from application of Central Florida PG to tilled land at levels up to 4 Mg ha-1 was minimal when compared to levels and variations of the natural background. Phosphogypsum-attributable additions were typically at or below the limits of detection by contemporary conventional radiation and radioactivity monitoring techniques. Thus, it is recommended that the restrictions on agricultural use be relaxed for application to tilled land.
With regard to relational data for future assessments, quantitative relationships were obscured because of the low level of radioactivity involved and the natural variability of the parameters being measured. Therefore, it was recommended that further studies incorporate higher treatment levels, additional replication, and modified procedures in order to improve the statistical power, provide measures of quantitative relationships, and describe time-dependent behavior. Such a study (Phase II) using a maximum rate of 20 Mg PG ha-1 is presently in progress.
In view of the minimal environmental effects of PG application to either established
bahiagrass pasture or tilled agricultural land cropped to annual ryegrass up to 4
Mg ha-1 of PG used in the study applied once in three years, it is recommended that
restrictions on the agricultural use of such PG at less than 4 Mg PG ha-1 be relaxed.
Application of PG with less than or equal to 23 pCi 226Ra g-1 at 2 Mg hal once every
three years or 0.65 Mg ha-1 annually would be comparable to the USEPA allowable rate
of 2,700 pounds PG acre-1 (3.024 Mg ha-1) applied biennially (1.51 Mg ha-1 annually)
for PG having less than or equal to 10 pCi 226Ra g-1.
FIPR Publication No. 01-085-127, Volume 1
Influence of Phosphogypsum on Forage Yield and Quality and on the Environment in Typical Florida Spodosol Soils. Volume I. Forage Yields and Quality of Bahiagrass and Annual Ryegrass Pastures Fertilized with Phosphogypsum as a Source of Sulfur and Calcium. University of Florida. July 1996.
Jack Rechcigl et al., University of Florida, IFAS. July 1996.