Progress report for GNE21-268
Project Information
- Rising sea levels, storms, and perigean spring tides push saltwater into coastal agricultural fields. This phenomenon, known as saltwater intrusion, alters nutrient cycling and damages crop yields. As sea levels continue to rise, saltwater intrusion will only worsen, with devastating consequences to agroecosystems along the coast of the Chesapeake Bay. Researchers and farmers alike are looking for solutions to adapt to and mitigate the effects of saltwater intrusion. Landowners may respond by altering their management practices. Farmers may 1) adapt by planting a salt-tolerant crop, 2) attempt to remediate soils with trap crops, 3) restore native marsh grasses, or 4) abandon fields altogether. My current project is to understand the survival of different crops and plant treatments under saltwater-intruded conditions and the potential for these plants to survive and to remove excess nutrients (e.g. sodium and phosphorus) from the soil, with the overall goal to benefit both the farming community and water quality in the Chesapeake Bay.
-
We have yet to understand the influence of these management practices on biogeochemical cycling, specifically P and salt (calcium, magnesium, potassium, sodium) dynamics. Improving soil health and plant productivity hinges on quantifying P and cation concentrations and pools in the soil, plant tissues, and porewater to identify differences in management options under SWI conditions. We established a study to investigate the concentrations and pools of P and cations (calcium, magnesium, potassium, sodium) in the soil and plant leaf tissue of four different management strategies represented by five proxy species: adapt (e.g., Cl-excluding soybeans [Glycine max], sorghum [Sorghum bicolor]), remediate (e.g., switchgrass [Panicum virgatum]), restore (e.g., saltmarsh hay [Spartina patens]), and abandon (e.g., weeds [Panicum dichotomiflorum]) on fields experiencing SWI.
-
We show that remediation and restoration management practices are efficient in taking up soil and porewater P that may ultimately decrease P runoff into the Chesapeake Bay. Overall, patterns of P pools were influenced by biomass levels, and after three years of planting, switchgrass and saltmarsh hay produced more biomass than weeds, sorghum, or salt-tolerant soybeans, and their production was influenced by the degree of SWI at each site. In sum, planting either switchgrass or saltmarsh hay would have economic and environmental benefits in mitigating the adverse effects of SWI. Additionally, we observed patterns of regulation in nutrient uptake in saltmarsh hay despite the imbalance of normal nutrient uptake due to saltwater Mg. Results from this study show restoration management practices are efficient at taking up soil Na, and may keep up with high Na concentrations and pools in the soil over time, which may benefit both soil and plant health on saltwater-intruded fields. Finally, we determined the best management practice to combat SLR and SWI (e.g., planting switchgrass or saltmarsh hay) will depend on the conditions of the land (e.g., high soil P or Na or both) and the goal of the landowner (e.g., economic or environmental benefit or both).
- Results from this study will help inform new management practices to increase soil health and maintain crop yields. By determining the efficacy of best management practices, we aim to help farmers keep their land profitable and remove excess P from the soil to promote better water quality. Improved economic and environmental sustainability in turn supports social sustainability by maintaining the viability of rural livelihoods and landscapes. Thus, it is important to understand the trajectories of different land management practices if farmers’ goals are to continue crop production while avoiding addition of excess nutrients to the Bay. The findings of this research will help inform farmer management decisions given economic and environmental objectives and tradeoffs. Finally, the goal of this work is to guide local best management practices and potential easement opportunities for landowners facing saltwater intrusion, and ultimately determine optimal strategies for climate resilience.
- To better understand P dynamics in soil and plant tissue (and how SWI effects soil chemistry and plant nutrient uptake), I had four objectives:
Obj.1.1. Quantify soil available P and total P, leaf tissue P, and porewater P concentrations and pools among adaptation, remediation, restoration, and abandonment systems and species
Obj.1.2. Determine if P biomass pools in representative species (e.g., adapt, remediate, restore, abandon) change along a SWI gradient
Obj.1.3. Determine if field management practices (e.g., adapt, remediate, restore, abandon) change soil P concentrations and pools along a SWI gradient
Obj.1.4. Determine if field management practices (e.g., adapt, remediate, restore, abandon) change porewater P concentrations along a SWI gradient
-
To better understand cation dynamics in soil and plant tissue (and how SWI effects soil chemistry and plant nutrient uptake), I had three objectives:
Obj.2.1. Quantify leaf tissue cation and soil available cation (calcium, magnesium, potassium, sodium) concentrations and pools among remediation, restoration, and abandonment systems and species
Obj.2.2. Determine if biomass cation pools in representative species change along a SWI gradient
Obj.2.3. Determine if field management practices change soil cation concentrations and pools along a SWI gradient
The purpose of this project is to identify best management practices for farmers facing saltwater intrusion (SWI). I will do this by determining the effects of adaptation, remediation, restoration, and abandonment treatments on soil and plant P and cation (e.g., Na) levels. These results will be used to determine the potential of each management practice to remove P and Na from soils to increase land productivity and reduce nutrient pollution of the Chesapeake Bay, contributing to the economic and environmental sustainability of systems in the region.
Climate change is often viewed as an impending threat. However, on the Eastern Seaboard of the U.S., SWI is already damaging agricultural systems, negatively impacting plant productivity and soil health [1]. Unfortunately, SWI will only worsen as sea levels continue to rise [2, 3]. Saltwater intrusion is the movement of sea salts into freshwater aquifers and shallow groundwater tables. Sea level is projected to rise by 0.2-0.5 meters by 2050 [4], concurrently with increased frequency of storm events and drought. Saltwater intrusion encroaches on coastal agricultural land connected to saline water bodies (e.g. tidal channels) through ditch networks [5]. This phenomenon causes an alteration in biogeochemical cycling, leading to eutrophication through the loss of nutrients (e.g. P), decreased land productivity through soil aggregate dispersion (e.g. due to Na), and overall loss of farmland through the flooding of fields and conversion of uplands into tidal marshes [3].
Agriculture is a way of life on the Eastern Shore of Maryland. Centuries of cultivation have drastically altered soil physical and chemical properties in the region. This history of land management is now clashing with the new threat of SWI. Soil P levels are high in the region due to decades of fertilizer applications in the form of poultry manure (known as “legacy P”) [6, 7]. At the same time, SWI is changing soil chemistry through the introduction of elements like Na. Sodium can lead to swelling and soil aggregate dispersion, reducing drainage and infiltration [8]. Saltwater intrusion may also release legacy P due to competition between phosphate (PO4) and sulfate for iron (Fe) on soil mineral surfaces [9]. As soils become anoxic, Fe is reduced and releases PO4 into solution. With SWI, the sulfate in saltwater competes with PO4 for Fe binding sites as the soils dry, leaving P in solution and prone to loss [10, 11].
My team has observed soil salinity levels ranging from 5 to 20 ppt in farm soils across the Lower Eastern Shore of Maryland [12]. However, most agronomic crops in the region do not tolerate salinity levels above 1-3 ppt [13, 14]. Elevated salinity levels have forced farmers to reconsider management practices. Landowners in coastal regions facing SWI have several choices: continue farming with salt-tolerant crops, remediate fields for a few years before returning to agriculture, restore the land to native marshes (which creates wildlife habitat that supports recreational activities like duck hunting), or abandon salt-damaged fields entirely [3].
Cooperators
- (Educator and Researcher)
- (Educator and Researcher)
Research
The following objectives will be accomplished via field experiments established in Somerset and Dorchester Counties, Maryland. Soil and plant samples will be processed and analyzed at the University of Maryland, College Park (UMCP).
Experimental design
I will leverage three experimental sites and one no-salt control site established by Drs. Tully, Gedan, and Miller (see Experience and Roles) on the Lower Eastern Shore of Maryland in 2018. The experiment was designed to examine the survival of different treatments under SWI. My research will add a new element - the potential for these treatments to remove Na and P. Thus, my work will add a novel element, which is highly relevant to the NE SARE objective of good stewardship, profitability, and quality of life for farmers and the agricultural community by guiding best management practices and reducing nutrient losses to the Bay.
Field sites are located on three saltwater-intruded agricultural fields planted with four replicate blocks of six treatments: (1-3) salt-tolerant crop rotation (salt-tolerant soybean, Glycine max; sorghum, Sorghum bicolor; and barley, Hordeum vulgare), (4) perennial remediation species (switchgrass, Panicum virgatum), (5) perennial restoration species (saltmarsh hay, Spartina patens), and (6) abandoned (natural recruitment control; native switchgrass, Panicum dichotomiflorum). Each plot is 3 m wide and 20 m long to capture a natural salinity gradient from the edge of the field (visible intrusion) to the center of the field (no visible intrusion). There is a 0.5 m buffer between treatments and a 1 m buffer between each block (Figure 1). The no-salt control field site has a layout of 16 plots with the same treatments (1-3) and a salt-tolerant crop rotation (salt-tolerant soybean, sorghum, and barley) with each entry point present every year, a remediation species (switchgrass) (treatment 4), and an abandonment species (native grasses) (treatment 6). As the field is not affected by saltwater, we did not install the restoration species (treatment 5; saltmarsh hay). Plots in the no-salt control are also 3 m x 20 m with a 0.5 m buffer between treatments and 1 m buffer between blocks. Soils at all sites are predominantly silt loams (Table 1) with little to no slope. The climate in the region is humid subtropical with a mean annual minimum temperature of 9.8 ºC, a mean annual maximum temperature of 20.6 ºC, and annual precipitation of 1270 mm [24, 25].

Table 1.1. Soil characteristics of field sites
Site |
Location |
Texture |
Baseline SOM |
Baseline bioavailable |
Clay (%) |
Silt (%) |
Sand (%) |
Taxonomic Classification |
Farm 1 |
Somerset Co., MD
Nearest town: Princess Anne |
Queponco silt loam |
2.1 ± 0.4 |
284 ± 73 |
16.74 |
38.31 |
44.95 |
mesic Typic Hapludults |
Farm 3 |
Somerset Co., MD
Nearest town: Crisfield |
Othello- Fallsington complex sandy loam |
2.7 ± 0.5 |
275 ± 16 |
8.50 |
21.83 |
69.83 |
mesic Typic Endoaquults |
Farm 6 |
Dorchester Co., MD
Nearest town: Cambridge |
Elkton silt loam |
2.3 ± 0.3 |
130 ± 25 |
23.57 |
71.59 |
4.84 |
mesic Typic Endoaquults |
LESREC (Control) |
Wicomico Co., MD
Nearest town: Quantico |
Mattapex silt loams |
2.0 ± 0.2 |
32 ± 8 |
22.93 |
64.75 |
12.32 |
mesic Aquic Hapludults |
Soil collection
In 2018, soil bulk density was collected from each block at Farm 6 and Farm 3 using a 5-cm diameter x 15-cm deep core and an AMS compact slide hammer (Core Sampler Complete, AMS, American Falls, ID, USA). Soil bulk density cores were collected every 0-10, 10-20, 20-30, and 30-60 cm. Soil samples were collected for bulk density in 2019 from each block at Farm 1 by digging pits to 55 cm and a 5-cm aluminum core was pushed horizontally into the side of the pit at 0-10, 10-20, 20-30, and 30-60 cm. In both cases, soils were returned to the University of Maryland - College Park and dried at 105ºC for 7 days. Core volume was used to calculate bulk density (g cm-3). Soil bulk density at LESREC was measured using the same process in September of 2021.
Baseline soils were collected from Farm 6 and Farm 3 in March of 2018 by compositing three cores taken at each depth (0-10 cm, 10-20 cm, 20-30 cm, and 30-60 cm) in each subplot using a 22-mm diameter push probe (AMS, Idaho Falls, ID, USA). Baseline soils were collected from Farm 1 and LESREC in November of 2018 using the same method as above. Baseline soils were analyzed for total C and N using dry combustion (LECO TruMac N, St. Joseph, MI). Soil available P, Ca, K, Mg, Na, and S were determined on an ICP following Mehlich-3 extraction (Mehlich 1984). Texture was determined using the hydrometer method (Gee and Bauder 1979; Table 1.1).
Starting in November 2019, we collected soil samples at the plot-level using the same depths as above (0-10, 10-20, 20-30, 30-60 cm; three replicates for each depth) at 5, 10, and 15 m from the edge of the field (Figure 1.2). Since the LESREC site did not have a salinity gradient, we sampled only at the plot-level. Because we found that soil Ca, K, Mg, Na, and P concentrations approached zero below a depth of 30 cm in 2018 and 2019, we only analyzed soils collected from 0-30 cm depths in 2021 and 2022.
Annual soil analysis
Annual soil samples were air-dried for at least three days before homogenization by grinding through a 2-mm mesh sieve. Ground soils were oven-dried for two days prior to analysis (60 ºC). Each depth was analyzed for texture (hydrometer method; Gee and Bauder 1979), total P (modified Kjeldahl digestion; Bradstreet 1954), and bioavailable P (Mehlich-3 extraction; Mehlich 1984). Soils from our 2021 collection were digested at 160 ºC with sulfuric/salicylic acid solution to break the carbon bonds between organic P compounds, which converts organic elements to a measurable inorganic form (e.g., total P) (Bradstreet 1954). The digestate solution was analyzed for total P on a LACHAT QuikChem (LACHAT Instruments Loveand, CO) using the molybdate-blue method for PO4-P (detection limit 0.01 mg PO4-P/L; Murphy & Riley 1962). Bioavailable P was also analyzed using the molybdate-blue method as above. Upon initial inspection, available P levels in the 30-60 cm depth were low (less than 25 mg kg-1) and similar across site, treatment, and section, so we focused on soil concentrations in the 0-30 cm depths for 2018 and 2021 collections.
Annual soil samples were air-dried for at least three days before homogenization by grinding through a 2-mm mesh sieve. Ground soils were oven-dried two days prior to analysis (60 ºC). Each depth was analyzed for texture (hydrometer method; Gee and Bauder 1979) and exchangeable Ca, K, Mg, and Na (Mehlich-3 extraction; Mehlich 1984). Exchangeable Ca, K, Mg, and Na in the 2021 soil samples were analyzed using air-acetylene flame Atomic Absorption Spectrometry (AAS; PerkinElmer Waltham, MA). Upon initial inspection, available Ca, K, Mg, and Na levels in the 30-60 cm depth were low and similar across site, treatment, and section, so we focused on soil cation concentrations in the 0-30 cm depths for 2018 and 2021 collections. We also found that there was not a significant difference between section across site, treatment, and depth, so we analyzed soil cation concentrations from the 0-5 m and 15-20 m distance from the ditch for the 2021 collection.
To calculate soil nutrient pools, the nutrient concentrations from each individual depth increment (e.g., 0-10, 10-20, 20-30, and 30-60 cm) was multiplied by the depth of the segment (in cm) and the bulk density (g cm-3).
Ground and dried soil samples were analyzed for electrical conductivity (EC) in 2018, 2019, 2020, and 2021 in a 1:5 soil to water slurry. Electrical conductivity was measured using a Thermo Scientific Orion Versa Star Pro (Thermo Fisher Scientific, Hampton, NH).
Plant collection
We used agronomic methods of biomass estimation for crops (adapt), and ecological methods for all other treatments (remediate, restore, abandon). Biomass was collected from the weeds, saltmarsh hay, and switchgrass plots from the three saltwater-intruded sites (Farm 1, Farm 3, Farm 6) in August 2018, 2019, 2020, and 2021. Biomass was collected from three replicate 0.25 m2 quadrats in four subplots: 0-5, 5-10, 10-15, and 15-20 m from the brackish water ditch that borders the study sites to capture the effects of salinity (Figure 1.2). Switchgrass was collected from LESREC at the plot level. Total biomass samples were clipped within 2 cm of the soil surface, weighed wet, allowed to air-dry, sorted to species, and then clipped again into sub-samples for processing. Sub-samples were oven-dried for at least 48 h at 60 ºC and weighed for conversion of total plant biomass to dry-weight-equivalent (DWE). Biomass of each species was summed at the quadrat-level and scaled to kg ha-1 to measure treatment effects on plant biomass. Each year, we determined the dominant weed species after processing. In 2018, fall panic grass (Panicum dichotomiflorum), a native variety of switchgrass, was a species common to nearly all weed plots across all sites therefore, was retained for tissue analysis. In the following years (2019-2022), the most common species was Digitaria sanguinalis, so this species was retained for tissue analysis.
Mature sorghum seed heads were collected from the sorghum plots in late September to early October and soybean plots in late October 2018, 2019, and 2020 along the saltwater intrusion gradient as described above (0-5, 5-10, 10-15, 15-20 m from the ditch). Soybean and sorghum yields were collected using a 0.25 m2 quadrat as a guide and collected by row in each plot. Five rows were collected of soybeans and three rows were collected of sorghum. Both sorghum and soybean seeds were passed through a plot combine (Massey-Ferguson 8XP Clarington, Ontario), bagged, and weighed on a scale to estimate the total yield per subplot (in bu acre-1). In the no-salt control field (LESREC), we collected biomass at the plot-level (as there was no salinity gradient) and processed as described above. Sorghum and soybean were collected in 2021 differently. Random individual sorghum and soybean plants were clipped within 2 cm of the soil surface from each subplot (three replicates) of their respective plots in October. The whole plant (including mature seed head or soybean seeds) was weighed wet, oven-dried for at least 48 h at 60 ºC, and weighed dry for conversion of total plant biomass to dry-weight-equivalent. The remaining crops were dried then passed through the plot combine, bagged, and weighed. Yields were adjusted to the correct moisture. Harvestable yields and biomass were converted to an area-basis (kg ha-1).
Plant analysis
Oven-dry (60 ºC) plant sub-samples from each treatment and year were homogenized by grinding through a 2-mm mesh screen (Wiley Mill, Swedesboro, New Jersey) and analyzed for total P. Plant tissue samples were digested as above (modified Kjeldahl; Bradstreet 1954). The digestate solution was analyzed for total P using colorimetry on a LACHAT QuikChem (LACHAT Instruments Loveand, CO) using the molybdate-blue method for PO4-P (detection limit 0.01 mg PO4-P/L; Murphy & Riley 1962). Biomass P pools, also known as P removal or P stocks, were calculated by multiplying nutrient concentrations by the aboveground biomass (DWE) and yields (DWE; kg ha-1) collected from the plots.
Oven-dry (60 ºC) plant sub-samples from each treatment and year were homogenized by grinding through a 2-mm mesh screen (Wiley Mill, Swedesboro, New Jersey). Only plant samples from the section closest to the ditch (0-5 m) and the section furthest from the ditch (15-20 m) were analyzed. Plant tissue and grains were pre-digested in concentrated nitric acid then digested in a heat block for 90 min after 30% hydrogen peroxide was added (Luh & Schulte 1985). National Institute of Standards and Technology (NIST) Standard Reference Material (SRM 1547 peach leaves) were included in every digestion to assure the effectiveness (95% confidence interval) of the nitric acid digestion. Digestion solutions were analyzed for total Ca, K, Mg, and Na on an ICP Emission Spectroscopy (Agilent Technologies Santa Clara, CA). Biomass Ca, K, Mg, and Na pools were calculated by multiplying nutrient concentrations by aboveground biomass (DWE) and yields (DWE; kg/ha) collected from the plots.
Porewater collection
Porous cup lysimeters (22 mm diameter; Soil Solution Access Tubes, Irrometer Riverside, California, USA) were installed in March 2018 (Farm 3, Farm 6) and October 2018 (Farm 1) to 60 cm depth at 5 m (visible saltwater intrusion) and 15 m from the edge of the plot (no visible saltwater intrusion). Lysimeters were installed by removing soil to 60 cm with a 22 mm diameter soil probe (AMS, Idaho Falls, ID, USA). Lysimeters were sealed at the soil surface with a bentonite/clay mixture (Benseal, Halliburton, TX, USA). Pilot studies confirmed that soil solution collection was only possible following rain events that were greater than or equal to 6 mm, thus soil solution was only collected following rain events of this level. Samples were collected at roughly 2-week intervals from March through November of each year (2018-2021). The day before sampling, lysimeters were purged of any water, and an internal pressure of -60 to -70 kPa was applied. Soil solution samples were collected, filtered (1 μm glass fiber filters), and stored in a freezer at -20°C until further analysis.
Porewater analysis
Electrical conductivity (EC) was measured using a Thermo Scientific Orion Versa Star Pro (Thermo Fisher Scientific, Hampton, New Hampshire, USA). A subsample of the filtered soil solution was acidified using 10% hydrochloride (HCl) solution (150 μL 10% HCl in 15 mL soil solution). This sample was analyzed colorimetrically on a LACHAT QuikChem (LACHAT Instruments, Loveland, CO, USA) using the molybdate-blue method for PO4-P (detection limit 0.01 mg PO4-P/L; Murphy & Riley 1962).
Statistical approach
A linear mixed-effects (LME) model (lme4 package for R; Bates et al. 2013) was used to examine the effect of management strategy (e.g., adapt, remediate, restore, abandon) on leaf tissue P concentrations and pools averaged at the plot-level. We examined each year and site individually with management strategy (treatment) and section (SWI) as the main effects and block as the random effect. If we found that there was no significant interaction between treatment and distance from the ditch, we analyzed the significance of treatment and section on leaf tissue P concentrations individually. Tukey post-hoc comparisons were used to evaluate significant differences among management strategies. Finally, univariate regression was used to determine if there is a relationship between soil and plant leaf tissue P concentrations.
A linear mixed-effects (LME) model (lme4 package for R; Bates et al. 2013) was used to examine the effect of management strategy (e.g., adapt, remediate, restore, abandon) on soil P concentrations and pools averaged at the plot-level. We examined each year and site individually with management strategy (treatment), section (distance from the ditch), and depth (0-10, 10-20, and 20-30 cm) as the main effects and block as the random effect. Once we determined there was a depth effect (and the interaction between the main effects were not significant), we examined each year, site, section, and depth individually (e.g. soil bioavailable P in 2021 from Farm 3 in S1 at 0-10 cm depth). Soil P concentrations were measured in 2018 (to understand baseline P levels) and 2021 (after a full crop rotation). Tukey post-hoc comparisons were used to determine significant differences among management strategies. Finally, univariate regression was used to determine if there is a relationship between topsoil EC (0-10 cm) and topsoil P concentrations.
A linear mixed-effects (LME) model (lme4 package for R; Bates et al. 2013) was used to examine the effect of management strategy (e.g., adapt, remediate, restore, abandon) on porewater P concentrations averaged at the plot-level. We examined each year and site individually with management strategy (treatment) and section (saltwater intrusion) as the main effects and block as the random effect. If we found that there was no significant interaction between treatment and distance from the ditch, we determined the significance of treatment and section on porewater P concentrations individually. Tukey post-hoc comparisons were used to evaluate significant differences among management strategies. Finally, univariate regression was used to determine if there is a relationship between topsoil EC (0-10 cm) and porewater P concentrations, soil and porewater P concentrations, and plant and porewater P concentrations. All statistics were run in the R environment for Mac (R Core Team 2021).
Correlation tests (PerformanceAnalytics package for R; Peterson et al. 2020) were performed to determine the relationship between different variables. We grouped all saltwater-intruded sites together and ran the test considering section as a variable for the plant leaf tissue and depth in addition to section as variables for the soil from year 1 to year 4. The first test was between plant leaf tissue cation concentrations (e.g., is saltmarsh hay Na levels influencing how much Ca is in their leaf tissue). The second test was between soil available cation concentrations (e.g., does the concentration of Na in the soil change the concentration of K in the soil). The final test was between plant leaf tissue cations and soil available cations (e.g., will increasing concentrations of Na in the soil effect the concentration of Ca in the leaf tissue).
A linear mixed-effects (LME) model (lme4 package for R; Bates et al. 2013) was used to examine the effect of species (e.g., switchgrass, saltmarsh hay, weeds) on soil exchangeable Ca, K, Mg, and Na concentrations and pools averaged at the plot-level. We examined each year and site individually with species (treatment), section, and depth (0-10, 10-20, and 20-30 cm) as the main effects and block as the random effect. Once we determined there was a depth effect (and the interaction between the main effects were not significant), we examined each year, site, section, and depth individually (e.g. soil bioavailable Na in 2021 from Farm 3 in S1 at 0-10 cm). Soil Ca, K, Mg, and Na concentrations were measured in 2018 (to understand baseline cation levels) and 2021 (after a full crop rotation). Tukey post-hoc comparisons were used to determine significant differences among treatments. Finally, univariate regression was used to determine if there is a relationship between topsoil EC (0-10 cm) and topsoil cation concentrations.
A linear mixed-effects (LME) model (lme4 package for R; Bates et al. 2013) was used to examine the effect of species (e.g., switchgrass, saltmarsh hay, weeds) on leaf tissue Ca, K, Mg, and Na concentrations and pools averaged at the plot-level. We examined each year and site individually with species (treatment) and section as the main effects and block as the random effect. Tukey post-hoc comparisons were used to determine significant differences among species. Finally, univariate regression was used to determine if there is a relationship between topsoil EC (0-10 cm) and plant leaf tissue cation concentrations. All statistics were run in the R environment for Mac (R Core Team 2021).
Chapter 1 Results
Plant leaf tissue P concentrations
We found similar patterns in plant leaf tissue P concentrations ([P]) among the three saltwater-intruded sites across all study years where the highest concentrations (3.8-4.7 g P kg-1) were always found in weed tissue and the lowest [P] were found in saltmarsh hay tissue (1.9-2.2 g P kg-1) with intermediate levels in switchgrass tissues (2.8-3.9 g P kg-1; p<0.03 in all cases). There was no clear relationship between plant leaf tissue [P] and distance from the salinity source although in some cases, we observed a significant effect of distance from the ditch. At the no-salt control site, we only planted switchgrass, which showed concentrations lower than the saltwater-intruded sites (2.7 g P kg-1).
Species biomass
Across all years, Farm 6 had the same patterns in biomass, where switchgrass accumulated the highest biomass levels (0.98-14.5 Mg ha-1), weeds produced the lowest biomass levels (0.11-3.36 Mg ha-1), and saltmarsh hay biomass had intermediate biomass levels (0.77-10.9 Mg ha-1; p<0.01 in all cases). The SWI gradient had an effect here in 2018 and 2019, where biomass increased with distance from the ditch (p<0.01 in both cases). In year 1 (2019) and year 2 (2020) at Farm 1, the weeds maintained high biomass levels (0.63-2.45 Mg ha-1), but there was no significant effect of distance from the ditch. We found similar patterns in 2021 at both Farm 1 and Farm 6, where switchgrass accumulated the highest biomass levels (4.77-11.3 Mg ha-1), weeds produced the lowest biomass levels (0.11-0.75 Mg ha-1), and saltmarsh hay biomass varied (1.72-8.55 Mg ha-1; p<0.01 in all cases). In contrast, saltmarsh hay at Farm 3 had the highest biomass in year 3 (2020) and year 4 (2021). At the no-salt control site, we only planted switchgrass, which had lower biomass than the SWI-impacted sites (0.18-6.78 Mg ha-1).
Plant leaf tissue P pools
In the first study year, weeds at all saltwater-intruded sites had the largest aboveground (AG) biomass P pools (p<0.003 in all cases). In years 2, 3, and 4, at Farm 3 and Farm 6 we observed the highest AG biomass P pools in the switchgrass plots (p<0.01 in all cases). However, in year 2 at Farm 1 weeds remained the highest AG biomass P pools (p<0.001). It was not until year 3 that switchgrass dominated, and we observed the highest AG biomass pools in switchgrass plots (p=0.01). P pools in the no-salt control showed a similar pattern to Farm 1 where switchgrass P pools were low in years 1 and 2, but increased in year 3.
Grain P concentrations
We found similar patterns in grain [P] among all saltwater-intruded sites across all study years with significantly higher [P] in soybean grain (4.5-9.0 g kg-1) than sorghum grains (5.1-7.0 g kg-1; p=0.01 in all cases; Figure 1.11). Soybean grain P concentrations were highest at Farm 1 in 2021 (6.1-8.0 g kg-1; p<0.001). There was no significant effect of distance from the salinity source at any of the sites each year (2019-2021). Finally, soybean P concentrations were lower (4.5-6.1 g kg-1) at the no-salt control site (LESREC) than the SWI-impacted sites in 2021.
Grain yield
The grain yield at Farm 1 in 2019 showed sorghum having higher yield (79.5-123.9 bu acre-1) than the county and no-salt control site averages (76 and 66 bu acre-1, respectively), but in the following two years, sorghum yield decreased to below average (0.78-23.5 bu acre-1). Soybean yield did not meet county and LESREC averages (55 and 58 bu acre-1, respectively) in any years at the saltwater-intruded sites (1.6-34.5 bu acre-1). Soybean grain was not collected at Farm 1 and Farm 3 in 2020 likely because of birds or deer eating the crops, so the values were zero. In 2021, the weeds at Farm 3 took over the field, and soybeans were unable to grow, and deer likely ate the grain at Farm 6.
Sorghum and soybean grain yields in bushels per acre from each research site on the Lower Eastern Shore of Maryland
from 2019 to 2021 compared to Dorchester and Somerset County (MD) averages calculated in 2021.
Year |
Farm |
Sorghum yield (bu acre-1) |
Soybean yield (bu acre-1) |
2019 |
1 |
79.5-123.9 |
0 |
2019 |
3 |
19.0-46.1 |
12.2-28.5 |
2019 |
6 |
31.1-43.5 |
17.2-27.5 |
2019 |
LESREC |
57.5 |
58.4 |
2020 |
1 |
0.8-4.8 |
0 |
2020 |
3 |
0.8-3.4 |
0 |
2020 |
6 |
1.3-10.6 |
2.1-7.8 |
2020 |
LESREC |
69.1 |
47.2 |
2021 |
1 |
12.0-23.5 |
1.6-34.5 |
2021 |
3 |
0.2-2.1 |
0 |
2021 |
6 |
2.2-8.8 |
0 |
2021 |
LESREC |
72.2 |
69.0 |
MD average |
|
76.0 |
55.0 |
Grain P pools
Sorghum and soybean grain P pools showed different patterns than grain yield and grain [P] at all SWI sites for all years. In year 2, sorghum grain P was slightly higher (2.71-9.81 kg ha-1) than soybean grain (3.46-5.95 kg ha-1; p=0.01), but there was no significant effect of distance from the ditch. There was no significant effect of treatment or distance from the ditch at Farm 3. Since there was only one species collected at Farm 1 and Farm 3 in year 3, there was no significant difference between species, but there was no significant effect of distance from the ditch either. Similarly, at Farm 6, there was no significance of distance from the ditch or treatment on grain P pools. The following year, there was only a significant difference between species at Farm 1 (p=0.01). There was a significant effect of distance from the ditch at Farm 3 (p=0.01) on grain P pools, but not at Farm 1 and Farm 6. Finally, the no-salt control (LESREC) soybean grain P pools were higher than pools at the SWI sites (22.5 kg grain-P ha-1).
Soil extractable P concentrations
In the soil, concentrations of Mehlich-3 extractable P ([M3P]) decreased by 30-50% in the topsoil (0-10 cm) at two sites after three years of planting (33.9% at Farm 3 and 52.8% at Farm 6). At Farm 6, [M3P] decreased from 109.3 mg kg-1 in 2018 to 51.6 mg kg-1 in 2021 in the top 10 cm of the soil. There was a significant effect of distance from the ditch at Farm 6 (p<0.05), where the section closest to the ditch (0-5 m) had higher [M3P] than the section furthest from the ditch (15-20 m). There was a significant effect of depth at all sites ([M3P] decreases with depth; p<0.01), but there was no effect of treatment.

Soil total P concentrations
Overall, soil total [P] decreased at all SWI sites after three years of planting, and concentrations decreased by 20% in the topsoil at Farm 1 and Farm 6 (19.6% and 18.3%, respectively). The total [P] in 2018 at Farm 6 was 507.4 mg kg-1 and decreased to 414.5 mg kg-1. There was a significant effect of distance from the ditch at Farm 3 and Farm 6 (p<0.05), where the section closest to the ditch (0-5 m) had higher [M3P] than the section furthest from the ditch (15-20 m) at Farm 6 and the opposite was seen at Farm 3. Similar to the soil extractable [P], there was no significant effect of treatment, and there was a significant effect of depth (total [P] decreases with depth; p<0.01).

Soil extractable P pools
Similar to the M3P concentrations, soil M3P pools decreased by 30-50% in the topsoil at two sites after three years of planting (33.9% at Farm 3, 52.9% at Farm 6). There was a significant effect of distance from the ditch at Farm 3 and Farm 6 (p<0.05), where the section closest to the ditch (0-5 m) had higher M3P levels than the section furthest from the ditch (15-20 m) at Farm 6 and the opposite was seen at Farm 3. There was a significant effect of depth (p<0.01). In 2018, soil M3P pools in the 0-10 cm depth (176-352 kg ha-1) had the highest M3P pools, the next 10 cm (10-20 cm) had intermediate levels of M3P pools (48.02-267.6 kg ha-1), and finally the deeper depth (20-30 cm) had the lowest M3P pools (8.47-77.4 kg ha-1). In 2021, soil M3P pools in the topsoil (0-10 cm) were greater (83.0-301.0 kg ha-1) than the 10-20 cm depth (19.9-150.0 kg ha-1) and the 20-30 cm depth (7.44-51.49 kg ha-1).

Soil total P pools
Following a similar pattern as the soil total P concentrations, soil P pools decreased, but by 20% in the topsoil at two sites after three years of planting (18% at Farm 1, 18.4% at Farm 6). There was a significant effect of distance from the ditch at all SWI sites (p<0.05), where the section furthest from the ditch (15-20 m) had higher P levels than the section closest to the ditch (0-5 m) at Farm 1 and Farm 3 and the opposite was seen at Farm 6. There was a significant effect of depth (p<0.01), where baseline total soil P pools in the 0-10 cm depth (819-993 kg ha-1) had the highest total P pools, the next 10 cm (10-20 cm) had intermediate levels of P (441-889 kg ha-1), and finally the deeper depth (20-30 cm) had the smallest total P pools (247-503 kg ha-1). After three years of planting, the depth effect remained consistent but total soil P was reduced at all levels; total soil P pools in the topsoil (0-10 cm) were greater (668-834 kg ha-1) than the 10-20 cm depth (326-743 kg ha-1) and the 20-30 cm depth (223-432 kg ha-1).

Porewater P concentrations
In the first study year, porewater soluble reactive P ([SRP]) was highest in the summer (0.047-0.057 mg L-1) and decreased after planting (0.026-0.036 mg L-1). In 2019, porewater [SRP] decreased slightly (0.019-0.022 mg L-1), but was still high in the summer months (0.034-0.058 mg L-1). In the summer of 2018 and 2019, porewater [SRP] was highest under the saltmarsh hay plots (0.057-0.058 mg L-1) and lowest under the switchgrass plots (0.034-0.047 mg L-1). The last lysimeter collection occurred two years after planting (in 2020), and porewater [SRP] decreased significantly under each treatment in both the summer and autumn ([SRP] < 0.01 mg L-1; p < 0.05), although there were no significant differences between treatments or between distance from the ditch at any of the SWI sites.

Chapter 1 Discussion
Our results have implications for how planting remediation and restoration species can draw down legacy soil P in coastal agricultural soils experiencing SWI. After analyzing the aboveground biomass and soil beneath five proxy species for management practices (adapt, remediate, restore, abandon) on three farms experiencing SWI, we were able to determine if soil and biomass P pools changed along a SWI gradient. We show that remediation and restoration management practices are efficient in taking up soil and porewater P that may ultimately decrease P runoff into the Chesapeake Bay. Overall, we saw significant differences in plant and soil P pools among each of the SWI sites, with relatively consistent trends across years (Figs. 1.5, 1.6, 1.7), which indicates that even with differences among our saltwater-intruded farm fields (degree of SWI, precipitation, soil type, etc.), patterns of P pools were influenced by biomass levels, and the uptake of P in biomass was consistent with the decrease in soil P and M3P concentrations and pools after three years of planting. Legacy P in our region due to historic application of poultry litter has led to high levels of both total and available P in the soil that could be mobilized and released when fields are exposed to saltwater (Weissman et al. 2019). Previous work showed that agricultural soils have a than wetlands when fields are allowed to flood with saltwater (Weissman et al. 2021), therefore it is important to determine best management practices that will draw down soil P. In this study, after three years of planting both remediation and restoration species (switchgrass and saltmarsh hay), there was an increase in aboveground biomass P pools as biomass accumulated over time, and this was associated with a decline in soil total and available P pools.
Switchgrass is efficient at taking up soil and porewater P
We found switchgrass biomass had higher P pools than both saltmarsh hay and weeds from saltwater-intruded fields, suggesting that switchgrass could be used as a P removal tool. Previous research has shown switchgrass can accumulate large quantities of biomass after the establishment year and can last over 5-10 years (McLaughlin & Kszos 2005; Wullschleger et al. 2010; Richner et al. 2014). We found that high levels of switchgrass biomass were associated with large aboveground P pools, and soil total and available P pools decreased after three years of planting (Figs. 1.5, 1.6, 1.7), which indicates that species with high levels of biomass and deep root systems can uptake high levels of soil P (Schmer et al. 2011; Basyal & Emery 2020). Porewater P was also lowest under the switchgrass plots after the second year and continued to decrease in the third year, and this was likely because of the decrease in soil P pools, so there is less P to be released as SRP (Meyerson et al. 1999). The difference between treatments on soil and porewater P were likely not significant because of switchgrass roots extending into nearby plots. Therefore, planting switchgrass on coastal agricultural fields could be a mitigation strategy to remove soil P, if harvested, and may ultimately decrease the amount of P available to runoff into the Chesapeake Bay (Schmer et al. 2011; Ashworth et al. 2017; Rivera-Chacon et al. 2022).
Switchgrass is a practical option for planting in the mid-Atlantic. First, switchgrass is more tolerant of saline and dry conditions due to its deep root system and strong associations with mycorrhizal fungi, so it is able to take up more water and nutrients than most agronomic crops (Di Virgilio et al. 2007; Schmer et al. 2011; Meyer et al. 2014; USDA NRCS 2020). Additionally, the deep, extensive root system can be as deep as three meters and as wide as six meters, allowing switchgrass to act as a buffer to take up soil N and P and prevent nutrient loss to nearby waterbodies (Blanco-Canqui et al. 2004; Lemus et al. 2008; Kibet et al. 2016; Kumar et al. 2019). As a perennial that does not need to be replanted each year with relatively low seed cost and no need for fertilizer amendments (UMD Extension 2010; AGMRC 2018), switchgrass incurs lower annual input costs than other crops. Finally, switchgrass can be used as a cellulose biofuel or poultry bedding, so landowners could continue to make a profit from planting, albeit at a reduced price per acre than typical corn and soybean crops ($265 per acre compared to $749 and $402 per acre, for corn and soybean respectively; Moyle et al. 2016; Purswell et al. 2020; UMD Extension 2022). In sum, this study showed switchgrass would provide a great tool to buffer against P loss, maintain low input costs, and be used as a biofuel or poultry bedding.
Saltmarsh hay is efficient at taking up P in soil with high salinity
We found saltmarsh hay biomass had intermediate biomass P pools compared to other treatments from saltwater-intruded fields, suggesting the potential to be used as a P removal tool. Previous research indicated that fields experiencing SWI will naturally transition to host more native marsh species such as saltmarsh hay (Gedan et al. 2019). Our work shows that the increase in saltmarsh hay biomass and aboveground P pools provide an opportunity to harvest this biomass to remove both sodium and P in the soil. At Farm 3 where salinity was the highest, saltmarsh hay had the highest biomass levels because of the higher tolerance to salt concentrations and was able to uptake similar levels of P as switchgrass (Figure 1.5). This uptake was associated with decreased levels of soil P pools and porewater [SRP] (Figs. 1.5, 1.6; Figure S1.7). Similar to switchgrass, our work indicates that allowing the natural conversion of fields to saltmarsh hay, or deliberately planting this species can serve as a tool for reducing P losses to nearby waterbodies by utilizing their high root porosity to maintain nutrient uptake (Burdick & Mendelssohn 1990; DeLaune et al. 2005).
Saltmarsh hay is a practical option for planting in high salinity soils on the Eastern Shore. First, saltmarsh hay is a halophytic C4 grass that can create a new wildlife habitat for waterfowl hunting and aid in restoring the land to native marsh. In addition to maintaining a profit by allowing hunting activity, landowners can receive financial assistance and guidance from state- and federal-level conservation programs that can support restoration practices (e.g., Conservation Reserve Enhancement Program; CREP 2022, Environmental Quality Incentives Program; EQIP 2022, Wetland Reserve Easements; WRE 2022). Finally, planting saltmarsh hay (Spartina patens) may facilitate the transition to tidal marshes that can protect coastlines from flooding by reducing strong winds and storm surges with their strong-stems and other physical properties (Hu et al. 2015; Leonardi et al. 2018). The restoration approach has been shown to save landowners in the Delaware, Maryland, Virginia (DMV) area between $4.5-23.8 million annually. For example, in Maryland, coastal wetlands reduced property damages from $20 million to $15.5 million from severe weather and flooding (Smith & Katz 2013; Narayan et al. 2017; USGCRP 2017; Li et al. 2020). Salt marsh presence reduces average annual flood losses by 18-70% (Narayan et al. 2017). In sum, restoring farm fields has benefits for nearby waterbodies by taking up soil P, landowners by maintaining revenue, and communities by protecting coastlines.
Grain crops will not maintain high yield with increasing SWI
The alternative grain crops tested in our field trials demonstrated efficient grain yield P uptake rates when the crops were able to survive. Sorghum is a salt-tolerant crop (relative to corn and soybeans) that maintains a good yield under increasing saline conditions (Grieve et al. 2012). Our data suggest that sorghum is efficient at taking up soil P pools, and we predict that soil P levels may decrease if the total aboveground biomass of the grain crop was removed at harvest. Utilizing this mitigation strategy would allow continued farming operations on the land because there is a market for sorghum as feed on the Lower Eastern Shore of MD (J. Miller, personal communication). Unfortunately, the value for the rest of the plant beyond the grain and the cost for specialized equipment to harvest the whole plant does not outweigh the option to grow alternative species (e.g., switchgrass, saltmarsh hay). Additionally, we planted Cl-excluding soybeans, which had lower yield than expected even under increasing saline conditions. Both sorghum and soybean total aboveground biomass is usually returned to the field as chaff during the harvesting process, so significant removal of P from the system is not likely with the adaptation management strategy. Finally, our research shows that these alternative grains, although efficient in taking up P, will not maintain high yield with increasing SWI.
Education & Outreach Activities and Participation Summary
Participation Summary:
I will target farmers who are experiencing saltwater intrusion at local farmer field days and workshops (e.g. The Maryland Sea Grant Coastal Farming Challenges Workshop), so they are able to make informed decisions on how to manage their land. I will reach landowners through Soil Conservation District events in Somerset, Dorchester, and Wicomico counties and work with farmers to generate applied research questions and effective land management practices. The annual research team Stakeholder Meetings are another outreach venue; the next one is planned for February 2022 in Princess Anne, MD. The annual stakeholder meetings help inform farmers, extension agents, and other agricultural professionals of our findings and seek input on new research directions. We also discuss the advantages and disadvantages of enrolling in easement programs or using alternative best management practices. I will target extension personnel and other farm advisors through the previously mentioned methods as well as by sharing findings at the Mid-Atlantic Crop Management School. I will also participate in a Mid-Atlantic SWI Conference (2023) that will include other regional (VA, NC, GA) research results from stakeholders. I will produce at least one Fact Sheet and at least two Extension short communications providing updates on the project. These will be disseminated to the farming community through the UMD extension network.
I will reach the scientific community through presentations at The American Society of Agronomy, the Crop Science Society of America, and the Soil Science Society of America (ASA-CSSA-SSSA) annual meeting in Baltimore, MD in 2022. I will share the results of my research with a large scientific audience through presentations at conferences and research symposiums and by publishing at least two articles in peer-reviewed journals such as Agriculture, Ecosystems, & Environment. This work builds on a larger project [funded by the National Institute of Food and Agriculture (NIFA)] that has an extensive outreach component. I will also leverage a new National Science Foundation Coastal Critical Zone Network (CZN) plan to disseminate my findings.
Finally, I will target policy-makers by sharing my findings with the Maryland Department of Agriculture, Maryland Department of Planning, and Maryland Department of Natural Resources, which are collaborating with Drs. Tully and Gedan to develop best management practices and potential easement strategies for salt-intruded lands. This proposal addresses pressing needs to better understand how agricultural landscapes will change in the coastal zone and how this research can inform current and future decision-makers and researchers. The knowledge we gain on Na and P uptake will inform best practices for adaptive farming, remediation, restoration, or abandonment of agricultural land impacted by SWI. For example, the Maryland Department of Natural Resources (DNR) is working to develop Coastal Resiliency pilot projects on salt-intruded farm fields to improve coastal communities’ resilience to the effects of climate change [30]. Our team has active collaborations with Maryland DNR, and I will present my results to the Director of the Office of Science and Stewardship and her team in an effort to inform policies that can support both farmers and the Chesapeake Bay.
Updates:
- University of Delaware Crop Management School Presentation - Recorded a 45-minute presentation on saltwater intrusion, how the soil and plants respond to saltwater intrusion (soil dispersion, sodium toxicity in plants), and possible new strategies to adapt to or mitigate the effects of saltwater intrusion (through the planting of salt-tolerant crops or transitioning fields to marshes) to farmers and extension agents.
- Research Recap Nutrient Management Sessions - Presenting for 45 minutes on saltwater intrusion, how the soil and plants respond to saltwater intrusion, possible new strategies to adapt to or mitigate the effects of saltwater intrusion, and some preliminary data. Talking with/answering questions for farmers about my research.
- Delmarva Soil Summit Presentation - Presenting for 15 minutes on the poster of my project with preliminary data for farmers, extension agents, and other researchers.
- Saltwater Intrusion Stakeholder Meeting Presentation - Presenting for 15 minutes on saltwater intrusion and my research results. Talking with/answering questions for farmers about my research.
- ASA, CSSA, SSSA (Tri-Societies) Annual Meeting Presentation - Presenting on my research project main findings titled Management Options for Farmers Facing Saltwater Intrusion on the Eastern Shore of the Chesapeake Bay for 15 minutes. Talking with/answering questions for other researchers and farmers about my research.
- ASA, CSSA, SSSA (Tri-Societies) Annual Meeting Poster Presentation - Presenting for 5 minutes on the highlights of my poster titled Improving Soil Health and Plant Productivity on Coastal Agricultural Fields Facing Saltwater Intrusion on the Eastern Shore of the Chesapeake Bay and talking with/answering questions for other researchers and farmers about my research.