Applications of Gypsum and Ammonium Sulfate Change Soil Chemical Properties of a Salt-Affected Agricultural Soil

Irrigation water with high electrical conductivity (EC) compromises the sustainability of agricultural soils. Calcium sulfate (CS) or gypsum is commonly used on removal of soluble ions such as sodium (Na), however, large applications of CS can affect soil pH, EC, and nutrient availability to plants. The objective of this study was to investigate the effects of CS and ammonium sulfate (AS) rates on the soil pH, EC, and exchangeable cations in a salt-affected agricultural soil. Samples from the 0-20 cm soil depth layer were collected from an agricultural soil reported to have low potato yield due to high EC irrigation water. Soil was incubated with rates ranging from 0 to 4000 kg ha-1 of CS and 0 to 600 kg ha-1 of nitrogen (N) using AS. The treated soil was incubated for 60 d at 25 ºC and moisture was maintained at 60% of soil field capacity. After incubation, the soil was analyzed for pH, EC, Na, manganese (Mn), and zinc (Zn). Increasing rates of CS resulted in a small decrease in soil pH and a significant linear increase in soil EC, while the application of AS linearly reduced the soil pH and quadratically increased soil EC. The application rate of 200 kg ha-1 of N as AS resulted in a decrease of soil pH from 5.9 to 5.2, while the EC increased from 1.3 to 3.0 dS m-1. Extractable Na increased linearly with the application of AS due to its effect on the soil pH. The soil extractable Mn and Zn were not affected by the application of CS. Applications of AS resulted in a linear increase in soil extractable Mn and Zn concentrations, respectively. Results from this incubation study suggest that the use of large rates of CS for consecutive years may further impair soil conditions for cropping in areas with high EC in the irrigation water.


Introduction
Groundwater is an important natural resource for domestic, agricultural, industrial, and public use in Florida as well as other parts of the world. The soils in northeast Florida were formed primarily from marine deposits and the proximity of the coastal areas leads the Floridian Aquifer System to be frequently intruded with saltwater (Frazee Jr. & McClaugherty, 1979). Saltwater intrusion is defined as the movement of saline water into freshwater aquifers, which can lead to contamination of freshwater sources. Saltwater intrusion occurs naturally to some degree in most coastal aquifers, owing to the hydraulic connection between groundwater and sea water. Because saltwater has a higher mineral content than fresh water, it is denser and has a higher water pressure with high EC. Activities like intensive groundwater pumping for public, industrial, and agricultural consumption can exacerbate saltwater intrusion (Barlow, 2003;Xiao et al., 2019). Particularly in northern St. Johns County, in Florida, the water type from the Upper Floridian Aquifer has high concentrations of calcium (Ca 2+ ), magnesium (Mg 2+ ) and sulfate (SO 4 2¯) and the presence of sodium chloride (NaCl) represents a mixing of saltwater with fresh aquifer water (Spechler, 1994).
The counties of St. Johns, Putnam and Flagler in northeast Florida are important agricultural production areas of winter/spring vegetables and potato (Solanum tuberosum L.). Seepage irrigation, also called subirrigation, has been the predominant irrigation method for vegetable production in Florida (Dukes et al., 2010). The presence of a shallow impermeable soil layer between 1 to 3 m below the soil surface allows growers to control the water table level for irrigation by pumping fresh groundwater in the fields. A series of furrows and ditches assist with the water distribution. A significant volume of groundwater is required to raise and maintain the water table level just below the plant root zone (Ferreira et al., 2017;Liao et al., 2016). After rainfall, generally the water table level is receded by draining the excess water offsite (da Silva et al., 2018).
Saltwater intrusion has severely impacted agriculture in northeast Florida. In recent years, there has been an increase in salt content in the irrigation wells, which might be attributed to low rainfall years and increasing water pumping for crop irrigation (Yarney, 2017). The problem is exacerbated with the use of seepage irrigation that applies large volumes of water with high salt content. Salts tend to accumulate in the soil profile above the impermeable soil layer, causing severe reduction in vegetable productivity in dry years. In the past 8 years, most of the growers have abandoned fertilizer broadcasting practices for more efficient application techniques such as banding of granular fertilizer or localized application of liquid fertilizers using knife applicator. Fertilizers with a high salt index can lead to an increase in soil EC with cations such as Ca 2+ , Mg 2+ and sodium (Na + ) and anions such as chloride (Cl¯), SO 4 2¯, bicarbonate (HCO 3¯) , and carbonate (CO 3 2¯) (Dunlop et al., 2019;Havlin et al., 1999;Mao et al., 2016;Vargas et al., 2015).
The use of CS on potato fields has become very popular among growers in recent years due to its low cost. In northeast Florida, applications of Ca are primarily used to minimize some potato tuber disorders like brown center, which are associated with Ca deficiency (Palta, 1996). Although the benefits of CS on tuber yield, tuber specific gravity, hollow heart or chip color were not confirmed to justify a routine application of CS on potato (Silva et al., 1991), CS has been periodically applied by growers. Liao et al. (2015) surveyed 32 private potato farms in northeast Florida and reported annual CS application rates between 896-2240 kg ha -1 , despite the fact that the soil Ca concentrations in the 0-20 cm soil depth layer on those same farms ranged between 374-3362 mg kg -1 , levels in which were classified as "very high" Liu et al., 2018). In other parts of the country severely affected by soil salinity, CS is used to promote the exchange of Na + by Ca 2+ in the soil (Mao et al., 2016). Once Na + is available in the soil solution, leaching practices can be applied to decrease the Na + content in the root zone (Sonon et al., 2015). However, in northeast Florida due to the shallow impermeable soil layer and high water table, leaching practices are less efficient than well-drained soils or areas with subsurface drain tile. Thus, management practices to prevent the introduction of salts in the field such as using low salt index fertilizers, avoiding excessive fertilizer application rates, and reducing irrigation water application have all become very important to maintain the sustainability of vegetable production in the region.
Sulfur (S) deficiency can occur in very sandy soils with low soil organic matter, especially following the continued use of sulfur-free fertilizer . The use of AS may be an alternative to supply S, and at same time, minimize N leaching compared to nitrate fertilizer sources. In fact, fertilizer blends used by potato growers may contain up to 24% of AS as a source of N. Although AS is a source for N and S for plants, AS can drastically reduce soil pH due to its acidifying power (Wang et al., 2018). Hart et al. (2013) stated that AS produces up to twice as much hydrogen (H + ) in soil compared to urea. Another possible disadvantage in using this source of N is the increase in soil EC. The application of N as AS has the potential to increase soil EC by 40 to 80% more than the same amount of N applied as urea (Vargas et al., 2015). According to Machado et al. (2014), for each gram per liter of AS applied in the soil, there is a 2 dS m -1 linear increase in EC in the soil solution.
The increase in EC and salt levels in the soil solution by applying CS and AS has been previously documented (Dunlop et al., 2019;Havlin et al., 1999;Machado et al., 2014;Mao et al., 2016;Vargas et al., 2015). However, many of these studies did not investigate the implications of application rates of CS and AS on the exchangeable cations in soils with high Na + content or those subjected to saline water irrigation, as in northeast Florida. The goal of this study is to raise awareness to the fact that management practices related to application of CS and some sources of fertilizer can exacerbate the salinity problem in the region. More specifically, the objective was to evaluate the effect of increased rates of CS and AS on soil pH, EC, and exchangeable cations in an agricultural soil with a high content of Na + .

Selection of the Area of Study
The main part of the study consisted of a soil incubation experiment conducted using soil samples from a commercial potato farm located in Hastings, FL, United States. The 200 ha farm has been under cultivation for more than 50 years, and the potato yields have decreased in some areas of the farm. The soil on this farm was classified as sandy loamy, siliceous, active, hyperthermic Alfic Alaquods belonging to Wabasso Fine Sand Series (Readle, 1983). Three areas, with low, medium, and high relative potato yield, were selected by the grower. The fields were irrigated with seepage. Composite soil samples from approximately 50 subsamples from each area were collected from the 0-20 cm soil depth layer during potato tuber initiation in Mar. 2012. The soil samples were homogenized, sieved (2 mm), and air-dried. The soil chemical characterization of the three production areas is presented in Table 1. Table 1. Mean values of soil chemical characterization of 0-20 soil depth layer using Mehlich-1 and potato foliar analysis sampled from three different areas based on relative potato yield levels (low, medium and high). Samples were taken during tuber initiation at Hastings, FL, Mar. 2012  Note. z EC, soil electrical conductivity; soil pH in water; sodium (Na); phosphorus (P); potassium (K); calcium (Ca); magnesium (Mg); nitrogen (N). y critical value for macronutrients for potatoes at first blossom .
Thirty-five potato foliar tissue samples per area were collected following the procedure described by Stark and Westermann (2008) and sent to the University of Florida Analytical Research Laboratory in Gainesville, FL, for analysis of N, phosphorus (P), Ca, potassium (K), and Mg. Based on the results of soil chemical characterization and plant analysis, the area with a low relative potato yield and a higher soil content of Na was chosen for the incubation study. After the potato harvest in Apr. 2012, 90 soil samples were collected from the 0-20 cm soil depth layer in a 40 ha area of the low potato production area and combined into a large composite sample. The soil was sieved (2 mm), homogenized, and air-dried. The incubation study was initiated immediately after soil was dried. The chemical characterization of the soil used in the incubation study is presented in Table 2. Table 2. Mean values of the soil chemical characterization (Mehlich-1) of the 0-20 cm soil depth layer used in the incubation study collected from a "low potato yield level area" after tuber harvest at Hastings, FL, May 2012.

Soil Incubation with Calcium Sulfate and Ammonium Sulfate
The incubation of soil with CS and AS was conducted at the Horticultural Sciences Department of the University of Florida in Gainesville, FL. A subsample of 200 g of air-dried soil was incubated with 12 rates of CS (100, 200, 400, 600, 800, 1000, 1400, 1800, 2200, 2800, 3400, and 4000 kg ha -1 of CS) and 12 rates of AS (25,50,75,100,125,150,175,200,300,400,500, and 600 kg ha -1 of N) and a unamended (control) treatment for each amendment in a single factorial experiment. All reagents used in this study were analytical grade of calcium sulfate dihydrate (CaSO 4 ·2H 2 O) (98%) and ammonium sulfate [(NH 4 ) 2 SO 4 ] (99.5%) (Acros Organics, Thermo Fisher Scientific, Waltham, MA).
The CS and AS rate calculations were performed on a weight basis assuming a soil bulk density of 1 g cm -3 . The experiment was layout was a completely randomized factorial design with 2 amendments and 12 rate rates with four replications. The soil samples were mixed with the respective CS and AS rates and placed into 250 mL glass containers. In the region, CS is broadcasted and then incorporated into the soil using disk harrow implement. Sources of N used for granular N fertilizer blend include ammonium nitrate (NH 4 NO 3 ) and (NH 4 ) 2 SO 4 . Granular fertilizer blends are generally banded or broadcasted, then incorporated into the soil using disk harrow at pre-plant or planting. Sidedress application of N can be applied as granular or liquid on the side of the potato hill and followed by the hilling operation using disk bedders. During the incubation, distilled water was periodically applied to maintain soil moisture equivalent to 60% of field capacity (FC). The internal temperature of the incubator was maintained at 25 o C. After 60 days of incubation, soil samples were removed from the containers and oven dried at 105 o C until constant weight. Samples were analyzed for pH in water in a 1:2.5 soil: water ratio. EC was analyzed using the soil saturated paste extraction method in a 1:1 soil:water ratio (U.S. Salinity Laboratory, 1954). Soil samples were also analyzed for Na, K, Ca, Mg, Zn, and Mn according to Environmental Protection Agency (EPA) method 200.7 (US. EPA, 1994;Sparks et al., 1996) after extracting with Mehlich-1 (Mehlich, 1953).

Statistical Analysis
Data were analyzed using the PROC GLM and REG procedures of SAS (version 9.4; SAS Institute, Cary, NC) at the level of significance of P ≤ 0.05 for soil extractable nutrient after incubation. Pearson's linear correlations between all soil chemical attributes and CS and AS were performed using PROC CORR of SAS.

Results and Discussion
In the low potato yield area, the initial survey sampling taken during tuber initiation stage revealed a soil EC at 3.79 dS m -1 and 252.9 mg dm -3 of Na + , as well as high concentration of P, K, Ca, and Mg in the 0-20 cm soil depth layer (Table 1). The plant tissue analysis (Table 1) confirmed that the three areas with various yields did not present any deficiencies related to N, P, K, Ca and Mg according to the critical levels reported by Hochmuth et al. (2018), which indicates that the low potato yield was likely due to high soil EC and Na + in the 0-20 cm soil depth layer. However, the levels of Na + in the plant and chloride (Cl -) in soil and plant were not evaluated in this study. Chloride toxicity could also have a played a role on the tuber yield reduction if present in high levels, despite the fact that high concentrations of Na + in the shoots can be primary driver of growth inhibition on potato compare to Cl -(Hutsh et al., 2019). Overall, the values of nutrients and Na + from the initial sampling were higher compared to the soil sampling performed after the potato harvest (Table 2), which was used for the incubation study. The lower nutrient concentration and soil EC in the 0-20 cm soil depth layer in the second sampling was attributed to the 98 mm of precipitation that occurred between samplings, the crop nutrient removal and the dilution of nutrients due to the mixing of soil from layers below 20 cm during the mechanical harvest. Nevertheless, the low yield area chosen for the soil incubation study with CS and AS still exhibited a considerable concentration of Na + (127.9 mg dm -3 ), which according to Sonon et al. (2015) would require an application of 4200 kg ha -1 of CS to remove exchangeable Na + after irrigation with salt-free water.
The impacts of the increased application rates of CS on the soil pH after 60 days of incubation are presented in Figure 1A. Soil pH decreased from 6.2 to 5.9 during incubation for the treatments without any CS or AS amendment. The soil moisture was maintained in ideal conditions for microbial decomposition and mineralization of plant residues particles from harvest (< 2 mm) which can contribute to the decrease in soil pH during the incubation period. There was a slight linear decrease in soil pH from 5.9 to 5.6 with the increasing rates of CS. The decreased soil pH corresponded to 0.063 units of pH for each 1000 kg of CS. The application of CS and its subsequent dissociation in the soil solution led to the formation of ion pairs between SO 4 -2 with Ca +2 , Mg +2 , Na + and aluminum (Al 3+ ), forming CS, magnesium sulfate (MgSO 4 ), sodium sulfate (Na 2 SO 4 ) and aluminum sulfate (Al 2 (SO 4 + ) 3 ), respectively, increasing the soil EC. Thus, the higher content of Ca +2 replace Na + in the soil exchange sites (Zambrosi et al., 2007) as the application rates of CS increased resulting in a decline in soil pH (McTee et al., 2017;Qayyum et al., 2017;Vet et al., 2014).  Note. z ns, not significant; *, significant at P  0.05; **, significant at P  0.01; ***, significant at P < 0.001. y soil pH in water; soil electrical conductivity (EC, dS m -1 ); sodium (Na, mg dm -3 ); potassium (K, mg dm -3 ); zinc (Zn, mg dm -3 ); manganese (Mn, mg dm -3 ); calcium (Ca, cmol c dm -3 ); magnesium (Mg, cmol c dm -3 ); base saturation (V, %).
The soil EC is dependent on the concentration of electrolytes (Nadler & Frenkel, 1980) and this phenomenon corroborated with the linear increase in soil EC with the increasing rates of CS ( Figure 1A). The Ca 2+ in soil solution can replace the Na + in the exchange complex, forming salts with high solubility as Na 2 SO 4 and MgSO 4 , which would be prone to leaching if conditions permit (Ali et al., 2007). The increase in EC as a function of CS application corroborates with the study conducted by Chun et al. (2001). The authors applied four doses of FGD-gypsum, 0; 5800; 11600; and 23100 kg ha -1 in a saline-sodic soil and the corresponded increase in soil EC was 0.5; 0.75; 1.2 and 1.3 dS m -1 , respectively.
The effects of increased application rates of AS on the soil pH and EC are presented in Figure 1B. The AS has high acidifying potential (Wang et al., 2018), which would invariably reduce soil pH. In contrast to CS, the application of AS significantly reduced soil pH from 5.9 to near 4.6 resulting in a significant negative linear response ( Figure 1B). Under the soil moisture conditions of the incubation, the AS was dissociated and ammonium (NH 4 + ) became present in the soil solution. Under field conditions, NH 4 + may be absorbed by plants, or depending on the soil moisture and temperature conditions, it can be subject to nitrification (e.g., conversion of NH 4 + to nitrate (NO 3¯) ). However, the nitrification releases H + into the soil solution, which results in soil acidification (Hart et al., 2013;Havlin et al., 1999;Wang et al., 2018). In the present study under controlled conditions, the significant decrease in soil pH with the increased application rates of AS was largely due to the nitrification process, which also affected other soil chemical attributes, such as EC, base saturation, and exchangeable Na, Mg, Zn and Mn (Table 4). Note. z ns, not significant; *, significant at P  0.05; **, significant at P  0.01; ***, significant at P < 0.001. y soil pH in water; soil electrical conductivity (EC, dS m -1 ); sodium (Na, mg dm -3 ); potassium (K, mg dm -3 ); zinc (Zn, mg dm -3 ); manganese (Mn, mg dm -3 ); calcium (Ca, cmol c dm -3 ); magnesium (Mg, cmol c dm -3 ); base saturation (V, %).  ory, 1954) can the variation in trongly saline, pectively. In th ratically increa erately saline a rs is in the ran -1 causing up t of CS and AS a of CS, whereas re was a quadr as due to a rel valence to be e is classified a increase in C as been previou Ca with the dec uent Na leachi nt ( Figure 3B).