Saving the Citrus Industry: Self-Funded Agrivoltaic Shield Walls for Asian Citrus Psyllids

1. Introduction

The Florida citrus industry has declined sharply by over 90% in the last 20 years, falling from 300 million boxes in 2003–2004 to roughly 20 million boxes in 2023–2024 [1]. Florida had 208,183 total acres of citrus for the 2024-25 year, which was another drop of 24% compared to the 2024 growing season [2]. There are several factors explaining the collapse of the Florida citrus industry including: land cost pressure from development with the rising population, high production costs (e.g., increasing costs of fertilizers, pesticides, and irrigation needed for an increasing number of droughts), labor shortages, extreme weather (e.g., hurricanes like Ian in 2022 and Milton in 2024 as well as the 2025 freeze), but the primary component was citrus greening disease [1]. Citrus greening disease also called Huanglongbing (HLB) in the literature, was first spread to South Florida in August 2005 resulting in orange production decrease of 92.5% and grapefruit production decreased by 95.6% [3]. HLB is caused by the intracellular Candidatus Liberibacter species, which is primarily transmitted to the trees by Asian Diaphorina citri (D. citri) and African Trioza erytreae citrus psyllids [4,5]. Although both vectors are capable of transmitting several bacterial species [6], citrus greening disease in Florida is spread mainly by the Asian citrus psyllid (ACP), and this bacterial disease causes i) asymmetrical blotchy mottling and chlorosis of leaves, ii) premature fruit drop, iii) small, lopsided, poorly colored bitter fruit, and iv) progressive tree decline, reduced yield, and eventually tree death [7]. This is not just a problem for Florida. Citrus fruits are cultivated in over 140 countries, with an average yearly production of 161.8 million tons [8,9]. HLB is widely considered the most devastating disease the global citrus industries have ever faced [10]. There is no known cure for infected trees [11]. The primary three-part approach to controlling HLB has been removal of symptomatic trees from the grove, control of the psyllid vectors, and replanting with disease-free nursery stock [12]. In the most extreme situations, China has been effective at controlling the disease by region-wide comprehensive management (rogueing + vector control + replanting) [13]. Many other methods have been investigated to control the disease including trunk injections of oxytetracycline (OTC) that improve yield and fruit quality and in parallel individual protective covers (IPCs), which delay infection for young trees [14]. In addition, thermotherapy, antimicrobial peptides, and transgenic approaches are under investigation [15]. Even with all of these approaches, it is clear HLB is causing severe multi-billion dollar/year citriculture industry losses [16].

Abandoned citrus groves in Florida serve as reservoirs for D. citri, and they can fly long distances despite geographical barriers [17]. Fortunately, they generally fly only at the height of the tree canopies. Thus, one of the most successful approaches to controlling HLB is the use of physical barriers to prevent ACP from reaching the trees [18]. ACP prefer living on the edges of groves and so the border trees are the first to be colonized [19]. ACP barriers can be either living windbreaks (e.g., trees) or artificial screen barriers on the perimeter of groves. The use of living windbreaks significantly decreases the density of ACP on the border of citrus groves [20]. This windbreaks function primarily as physical barriers but may also provide a habitat for natural enemies of ACP [21]. The challenges of live windbreaks are numerous and include time to establish, cost, and their susceptibility to extreme weather (e.g., freezing). A large collaborative group of researchers including Sétamou (Texas A&M University-Kingsville), Martini, Diepenbrock, and Stelinski (University of Florida), Grafton-Cardwell and Rivera (University of California Riverside) are exploring border management strategies [19]. They recommend that screen barriers be built with a fence at least 3.7m (12 ft) high, 9-23m (30-40 ft) away from the border trees with psyllid resistant screens (e.g., high-density polyethylene (HDPE) plastic mesh) [19]. Experiments consisting of deployment of sticky cards at different heights revealed that most >99% of D. citri were collected on traps placed at or below 3.7 m (12 feet) along grove borders [22]. Thus a 3.7 m high fence, if completely encircling a grove, would be expected to effectively exclude the vast majority of psyllids. In a field study conducted in Texas [23], only south and east border fences of a grapefruit grove resulted in 55-98% reduction of ACP numbers relative to a control grove with no border fence. Even more aggressive is the newly introduced citrus under protective screen (CUPS) production systems, which are able to produce disease-free fruit by completely enclosing a grove in a screen house including the roof [24].

Physical barriers like fencing, screens and CUPS, unfortunately, all add to the already high production costs for citrus. This makes Florida citrus more likely to be displaced for more economic activities, including with large scale solar farms [25]. For example, CUPS significantly increases the cost of establishing or maintaining a citrus grove due to the high cost of screen-house construction, which averages $10.76/m² ($1/ft²) [24]. This amounts to $107,639 per hectare or $43,560/acre and makes economic citrus production very challenging as the gross revenue is $2,500 to $3,500/acre in a typical well-managed mature grove producing 300-400 boxes/year/acre; while production costs about $2,000/acre/year to maintain the trees [26]. Thus, if an acre is only making $500-$1,500/acre/year simple payback for CUPS is over 29 years [27]. What the Florida (and global) citrus industry needs is a free (or better yet profitable), method to erect physical barriers to prevent the ACP from infecting their fruit trees with HLB.

To provide that solution, this article investigates the technical, economic, and scaling potential of utilizing vertical solar photovoltaic walls to surround citrus groves and provide physical barriers to ACP. The solar photovoltaics (PV) in the agrivoltaic shield wall are profitable alone, and this solar electricity profit can be used to subsidize the entire barrier to the ACP. Vertical solar fencing is already in use in agrivoltaics, and only minor modifications would be necessary to make them ACP-tight. First, the physical requirements are detailed. Next, simulations are provided for a case study in Vero Beach Florida for the economic output of such an array. Finally, GIS is used to determine the potential impact on Florida’s electric grid if this practice was adopted state-wide.

2. Methods

2.1. Physical Requirements

The structural designs of the proposed agrivoltaics citrus shield fences are compliant with applicable regulations and standards, which include the Florida Building Code (FBC) and American Society of Civil Engineers (ASCE) guidelines. The following steps detail the design of the agrivoltaics citrus shield fence:

• Design Wind Speed and Risk Category

The vertical PV walls are assigned Risk Category 1 as per FBC Table 1604.5 since the facility is for isolated agricultural use [28]. As Vero Beach is in India River County, the design wind speed (V) is set to 150 mph (refer to FBC Figure 1609.3(1)) [28]. Moreover, following ASCE 7-22 Table 26.6-1 for solid freestanding walls, the wind directionality factor (K_d) is considered 0.85 [29].

2.

Velocity Pressure (q_z)

The PV fences are assigned exposure category C (open terrain) considering FBC Section 1609.4.3 [28]. For a height of up to 15 ft, the velocity pressure coefficient (K_z) is considered 0.85 as per ASCE 7-22 Table 26.10-1 [30]. The value of topographic factor (K_zt) is 1.0 for flat terrain following ASCE 7-22 Section 26.8 [31]. Lastly, the ground elevation factor (K_e) is to be kept 1.0 referring to ASCE 7-22 Section 26.9 [32]. The velocity pressure (q_z) is given by:

q_z=0.00256.K_z.K_zt.K_d.K_e.V² [kPa]

(1)

3.

Wind Load and Bending Moment

To determine the net pressure coefficient or force coefficient (C_n), the aspect ratio is ascertained using the length of the fence used in the case study (363m or 280m) and the fence height (1.134+1.134=2.268m). Since the aspect ratio is greater than 45, is considered 1.3 by assuming Case A for wind load [33]. Following ASCE 7-22 Section 26.11, gust factor (G) of 0.85 is used in the calculations [34]. The design wind load (P) is then calculated using the following equation:

W or P=q_z.G.C_n [kPa] (2)

The dead weight (D) of the module is considered to be 0.24 kPa [35]. Using FBC Section 1605.1, the load combination for allowable stress is ascertained as follows:

Load combination=1.0D+0.6W(or P) [kPa] (3)

The distributed wind load (w) is determined by multiplying the load combination by the tributary width (w_Tributary , which is the width of the module).

w=Load combination x w_Tributary [kN/m] (4)

The total force (F) is calculated by multiplying w with the height of the structure (which is two times the height of the module – 1.134+1.134).

F=wh [kN] (5)

The maximum moment is calculated using the following equation:

M=Fh/2 [kN.m] (6)

4.

Structural Member Selection and Deflection

The allowable bending stress (F_b) for A-36 steel is 160,000 kN/m² [36].

The material selection was performed by calculating the section modulus and satisfying:

S_required > M_max/F_b [kN.m] (7)

Actual deflection is estimated using the following equation:

Δ=wh⁴ /8EI [mm] (8)

It is compared with the deflection limit (L/120) where L is equal to the value of h.

5.

Foundation Design

The foundation is designed considering FBC Section 1806.2 for sandy soil (Class 5) with allowable lateral bearing pressure is 100 psf per foot of depth (15.7 kN/m² per m) [37]. The following formula is used for embedment calculation:

d=0.5A{1+[a+(4.36h/A]^0.5} [m] (9)

and where

A=2.34P/(S₁.b) [m²] (10)

where b=0.45, b is the diameter of round post or footing, d is the depth of embedment in earth, h is the distance in feet from ground surface to point of application of applied lateral force, P is the applied lateral force in pounds, and S₁ is the allowable lateral soil-bearing pressure.

2.2. Simulations and Case Study

A case study was conducted using the Florida Research Center for Agricultural Sustainability. It has 35 acres of citrus groves in the heart of Florida's Indian River Citrus District in Vero Beach, Florida [38] and represents the size of a typical small family farmer-based grove. A satellite image of the grove is shown in Figure 1 with the approximate locations and lengths of agrivoltaic shield walls shown in orange.

For the agrivoltaic setup, two different case study designs were considered.

First, the minimal agrivoltaic setup needed to provide an external barrier to the grove in Case 1 includes setting up fence PV along the perimeter of the Florida Research Center. The North-South (N-S) fences have lengths of 363 m, while East-West (E-W) fences have lengths of 280m each (see Figure 1).

BiHiKu6 CSW-550MB-AG PV modules by CSI Solar Co, Ltd. were selected for the simulations [39]. The dimensions of the module are 2,266mm x 1,134mm. The modules are considered to be installed in portrait layout as shown in Figure 2.

The above fence design consists of two 0.12m x 0.12m fence posts set in concrete footings 2.850m deep. Appendix A contains detailed technical drawings of the fence units. Citrus netting was placed above the panels at a height of 3.7m from the ground. All openings around the PV modules were sealed using addition citrus netting. This design allows the fence to be hermetically sealed while allowing maximum light penetration to the PV modules.

In Case 2, along with the exterior fence, two additional rows of PV wind breaks are also installed in N-S direction with lengths of 343m (considering 10m space on either side for the turning radius of farm equipment). This represents a case of maximizing solar electricity production while still enabling conventional citrus industry equipment to navigate the agrivoltaics grove.

Energy analysis is performed using the open-source System Advisor Model (SAM) [40]. The number of modules along the fence is calculated by dividing the total length of fences with the width of the module (shorter side of the modules since the modules will be installed in portrait orientation). Based on the number of modules and the power rating of each module, the total installed capacity is ascertained (Figure 3).

The input parameters for SAM are detailed in Table 1.

2.3. System Design Economics

The technical and economic viability of deploying a grid-tied PV system is evaluated by outlining the applicable Florida Power & Light (FPL) interconnection rules, rate structures, and available incentives.

Case 1 and Case 2 systems in Table 2 were evaluated exclusively under three tariff frameworks: the Public Utility Regulatory Policies Act of 1978 (PURPA) [43], Qualifying Facility framework [44], and FPL's purchase tariff schedules [45,46,47,48].

Economic analysis includes two scenarios. Scenario A analyzed net metered systems as if the facility matched the output of the PV with almost constant co-located loads e.g., with easily located computing or controlled environmental agriculture. In Scenario B, the facility has no on-site electricity load to offset. All generation is exported directly to FPL under the COG-1 as-available energy tariff at the avoided cost rate of 3.353 cents per kWh [47,48]. To bound the sensitivity of the agrivoltaic economics to capital cost, three CapEx sub-cases are evaluated for each combination of Case (1 or 2) and Scenario (A or B), yielding twelve simulations in total using the PV System Cost Model (PVSCM) Cost Benchmarks [49]. Sub-case 1, the worst-current CapEx, applies the published Modeled Market Price (MMP) of $1,712.12/kW-DC for Case 1 and $1623.48/kW-DC for Case 2, which reflects the price a third-party developer would charge an external customer in the current U.S. market with all eight PVSCM cost categories included (modules, inverters, structural balance of system, electrical balance of system, on-site fieldwork, off-site office work, and developer indirect costs). The operation and maintenance costs (O&M)1 for Case 1 and Case 2 are 29.12 $/kW.yr and 23.91 $/kW.yr, respectively. Sub-case 2 is the breakeven capital cost, defined as the $/W at which the bifacial PV yields zero net profit over the 25-year horizon. Sub-case 3 represents an owner-developed deployment on land already owned by the citrus operator, in which the four sub-elements within the PVSCM “Other” category attributable to a third-party developer entity are removed: sales tax on photovoltaic hardware, exempt under Florida Statute Section 212.08(7)(hh), which provides a permanent state sales and use tax exemption for solar energy systems and their components as certified by the Florida Solar Energy Center [51]; developer profit margin (no third-party markup, as the project is build-to-own); developer management overhead (absorbed within the existing farm operation rather than billed as a separate line item); and developer contingency reserve (the owner bears project risk directly). Unavoidable indirect costs that remain regardless of developer identity are retained at their benchmark values, including permitting, inspection, and interconnection fees, general liability insurance required by Florida Public Service Commission Rule 25-6.065(6)(e) [52], and financing transaction costs. The resulting Sub-case 3 MMP, computed directly from the PVSCM 2024 workbook with these sub-elements zeroed, is $1,444.73/kW-DC for Case 1 and $1,367.44/kW-DC for Case 2. Considering the fact that sub-case 3 is grower owned facility, O&M management fees were excluded in PVSCM model, resulting in O&M costs of 14.9 $/kW.yr for Case 1 and 14.64 $/kW.yr for Case 2.

Energy system configurations were simulated on SAMA [53,54] over a 25-year project lifetime using a rate escalation CAGR of 2% per year [55], and a nominal discount rate of 3.75% [56]. The interconnection rules for Florida are summarized in Appendix B. The bi-facial PV module was modeled into SAMA using pvlib.bifacial.infinite_sheds module [57].

Customer-owned renewable generation within FPL's service territory may interconnect under Florida's net-metering framework (FPSC Rule 25-6.065 [52]) or sell electricity to FPL under two tariff pathways: COG-1 (As-Available Energy, Sheet 10.100) and QS-2 (Firm Capacity and Energy, Sheet 10.300) [48]. Both agrivoltaic cases (529.2 kW AC and 811.6 kW AC) qualify as Tier 3 systems under FPL Tariff Section 9 [46] and, given their sizing at approximately 100% of modelled annual load, are well within the 115% annual consumption cap for net metering [58]. As summarized in Table 3, the three pathways offer distinct compensation structures: under net metering, excess generation exported to the grid is credited at the full GSD-1 retail energy rate (approximately 6.31 cents/kWh [47]), with any unused year-end credits paid out at FPL's COG-1 avoided-cost rate of 3.353 cents/kWh at secondary delivery voltage [48]; under COG-1, all generation is compensated purely at this avoided-cost rate, derived from FPL's fuel charge of 3.201 cents/kWh [47] plus a fixed variable O&M adder of 0.01540 cents/kWh adjusted by a secondary voltage multiplier of 1.0413 [48], with no capacity payments; and under QS-2, capacity payments are added but require firm delivery commitments (94% availability) that a solar-only system cannot reliably meet without battery storage, as noted under Section 366.91(3), Florida Statutes [59]. Given the high capital costs associated with battery storage systems [60] and their adverse impact on project economics, QS-2 was not analyzed in this study. Detailed information on all three pathways and the associated interconnection requirements is available in Appendix C & Appendix D.

2.4. Scaling Agrivoltaic Shield Deployment Throughout Florida

The scaling potential of agrivoltaic shield walls for mitigating ACP intrusion in Florida was explored by adapting a previously developed GIS-based fence-mounted PV simulation framework [61]. The framework is applied here to quantify deployable PV capacity and energy generation along citrus grove perimeters. This is then expanded to other farms in Florida to analyze fence-mounted PV as functional biosecurity infrastructure.

2.4.1. Citrus Grove Identification and Spatial Scope

Crop sequence boundary shapefiles for the year 2023 were obtained from the United States Department of Agriculture National Agriculture Statistics Service (USDA-NASS) [62], and the Florida state boundary data were extracted. The resulting dataset was filtered for citrus and oranges using the land cover code attributed to citrus (72) and oranges (212) located in the metadata file provided by USDA-NASS. This process outputs polygons representing the active citrus and orange groves that will be referred to as citrus dataset in the remainder of the manuscript. The citrus polygons dataset was processed to remove inner holes (i.e. farm roads, barns, etc.) from polygons and dissolve adjacent farms to avoid double-counting common edges in accordance with the methodology. The process is repeated for all farms across Florida to quantify the total farm-fence mounted PV potential across the state.

2.4.2. Fence Geometry Extraction, PV Capacity Estimation, and Annual Energy Simulation

External boundary information, including fence perimeter length and orientation, was extracted for each group of farms resulting from polygons dissolution. The total PV capacity is calculated using the length of the fences as shown in equation (11).

$$P_{farm}=P_{mod}\sum _i^{N_{perim}}{\displaystyle \frac{L_i}{w_{PV}}}$$

(11)

where $P_{farm}$ (W) is the installable capacity on a farm, $P_{mod}$ (W) is the DC rating of a module, $N_{perim}$ is the number of perimeters on a group of farms polygon representation, $L_i$ (m) is the length of perimeter segment $i$, and $w_{PV}$ is the width of the PV module along the fence lengths. Each fence perimeter is considered entirely covered by PV modules in a continuous arrangement with no spacing between modules. A portrait layout configuration, matching the design in Figure 2, was analyzed where the PV modules are arranged vertically.

The orientation of each fence segment was incorporated into the energy simulation. Segment-level azimuth angles, combined with location-specific solar resource data, were used to estimate energy production for each boundary segment. Annual energy generation was computed at the segment level and aggregated to the farm, county, and state levels.

Meteorological data were obtained from the National Solar Radiation Database (NSRDB). County centroids were used to extract representative weather data, under the assumption that solar resource variability within individual counties is limited relative to the spatial scale of the analysis.

3. Results

3.1. Case Studies

3.1.1. Structural Analysis

The velocity pressure is calculated to be 1.99 kPa following Equation 1. Using Equation 2, the calculated value of design wind load was 2.20 kPa. The load combination resulted in a total load of 1.56 kPa while the value of distributed wind load is determined to be 3.53 kN/m. The maximum moment is found out to be 9.08 kN.m. The value of M_max/F_b is calculated to be 56.8. The section modulus for 120x120x5mm is 84.6. The actual deflection (16.22mm) with 120x120x5mm came out to be within the deflection limit (18.9) and therefore, is considered for fence posts. Considering a diameter of 0.45m, the depth of the foundation is found to be 2.85 m

3.1.2. Energy Analysis

The energy analysis demonstrates that 320 modules can be installed on the N-S fence; with 640 modules total for both fences. Moreover, for the E-W fences, the module count comes out to be 492. The total installed capacity for the N-S and E-W fences is estimated to be 352 kW and 270.6 kW. The energy generation potential for the N-S and E-W fence is determined to be 1373 kWh/kW and 1099 kWh/kW. By using the energy output for 1-kW system and the total installed capacity, the total energy output from N-S and E-W fences comes out to be 483.30 MWh and 297.39 MWh. The total energy output for Case 1 is estimated as 780.69 MWh each year (Figure 4).

For Case 2, an additional 302 modules can be installed on each wind shield – therefore, 604 modules will be mounted on both wind shields. This would result in an installed capacity of 332.2 kW. In terms of energy generation potential, the wind shield will produce 456.11 MWh. The total energy generation potential for Case 2 is 1236.80 MWh/year.

3.2. Economics

3.2.1. Economic Implications of Large-Scale Solar PV (Scenario A: Net Metering)

Scenario A models the agrivoltaics system under FPL's net-metering tariff, with the facility supplying its own electrical load and exchanging surplus and deficit kWh with the grid through monthly netting. The simulation uses the GSD-1 rate schedule effective January 1, 2026 [47] (base charge $33.71 per month, maximum demand charge $12.70 per kW, energy charge 2.825 cents per kWh, environmental charge 0.286 cents per kWh, fuel charge 3.201 cents per kWh, conservation charge $0.49 per kW, capacity charge $0.16 per kW, storm protection charge $1.80 per kW, transition credit -$0.12 per kW). The annual property load is set equal to the annual PV generation (827,119 kWh for Case 1 and 1,300,072 kWh for Case 2), reflecting the policy intent of Rule 25-6.065 [52] and remaining well within FPL's 115 percent annual-consumption sizing guideline [58].

Table 4 summarizes the key outputs of Scenario A across the three CapEx sub-cases of Case 1 and Case 2; the principal economic limitation common to all sub-cases is the FPL demand-charge structure, which is applied to maximum gross import demand and is not offset by PV output during typical peak hours. Quantitatively, under Scenario A with the GSD-1 tariff, the 25-year present-value bill savings reach $1,268,750 (Case 1) and $1,971,600 (Case 2) for all three CapEx sub-cases. Whether these bill savings recover the capital and O&M investment over the 25-year horizon depends on the CapEx sub-case. At Sub-case 1 pricing, both cases are below breakeven with net profit of -$162,259 (Case 1) and -$37,957 (Case 2), and no payback within the lifetime; Case 2 is close to recovery even under this most-conservative CapEx assumption. The Sub-case 2 breakeven CapEx is $1,451.25/kW (Case 1) and $1,583.65/kW (Case 2), reached at 25-year payback. Under Sub-case 3 (owner-developed), both cases reach substantially positive economics: Case 1 achieves ROI of 20.34 percent and net profit of $182,817; Case 2 achieves ROI of 29.51 percent and net profit of $384,985. IRRs are -1.18 percent (Case 1) and -0.18 percent (Case 2) at Sub-case 1 pricing, moving to 1.44 percent and 2.04 percent respectively at Sub-case 3 pricing. The Sub-case 3 advantage over Sub-case 1 reflects both the reduced capital cost from removing developer-attributable indirect cost components and the reduced annual O&M from removing the management fees for the owner-operated configuration, while the demand-charge component of $15.03 per kW, applied to the full gross import peak (94.42 kW and 148.41 kW) and not reduced by PV, contributes $17,030 per year (Case 1) and $26,767 per year (Case 2) of unavoidable cost regardless of CapEx.

3.2.2. Economic Implications of Large-Scale Solar PV (Scenario B, Wholesale Export)

Scenario B models the same two physical systems under the assumption that no on-site electrical load exists, so the agrivoltaics facility delivers all generated AC energy to FPL under the COG-1 As-Available Energy tariff at the secondary-voltage rate of 3.353 cents per kWh. There are no retail electricity purchases, no demand charges, no minimum base charge, and no sales-tax considerations on the export side (COG-1 sales are wholesale and outside the retail sales-tax framework).

Table 5 summarizes the Scenario B outputs across the three CapEx sub-cases. The economics are uniformly unfavorable: at Sub-case 1 pricing both cases produce deeply negative ROIs; the Sub-case 2 breakeven CapEx collapses to $503.17/kW (Case 1) and $635.7/kW (Case 2), which is well below any credible market price for an agrivoltaic system; and even under Sub-case 3 (owner-developed) pricing both cases remain solidly negative. The structural reason Scenario B fails across all sub-cases is that COG-1 compensates exports at 3.353 cents per kWh, which is approximately 53% of the GSD-1 retail energy rate that net-metered exports recover. For this mechanism to work financially from an energy-economics perspective, citrus grove owners therefore need to couple agrivoltaic shields to a co-located load matching their annual PV output (to qualify for net metering) and adopt the owner-developer deployment structure.

The two scenarios produce markedly different project economics, and the divergence is driven almost entirely by the effective per-kWh compensation rate for exported energy, not by differences in PV system cost, generation, or technical performance. Under Scenario A, exports are credited against retail imports at the GSD-1 retail energy rate of approximately 6.312 cents per kWh (excluding the demand-charge component, which is not netted), with only the residual December true-up balance cashed at the COG-1 wholesale rate of 3.353 cents per kWh. Under Scenario B, all exports are paid at COG-1 from the first kWh, with no monthly netting against retail charges.

3.3. Citrus and Statewide Farms Fence-Mounted PV Potential in Florida

According to the data illustrated in Figure 5, 32 counties in Florida have citrus farms in 2023. The average fence-mounted PV capacity across these counties is 0.4 GW with a standard deviation of 0.6 GW, while the average annual energy production is 0.5 TWh with a spread of 0.8 TWh, for the investigated portrait layout. This large dispersion observed across the data is the result of the farms number discrepancy between counties. Polk County, with 2,462 farms after polygons dissolution, exhibits the highest PV capacity with 2.5 GW (3.3 TWh) followed by Highlands and Hardee counties, with 1.5 GW (1.9 TWh) and 1.4 GW (1.8 TWh), respectively. On the other hand, the counties with a single citrus farm displayed the lowest capacity, including Gadsden and Palm Beach counties each with fence-mounted PV capacity of 0.2 MW and 0.4 MW, respectively. These capacities correspond to annual energy potential of 0.3 GWh and 0.5 GWh.

The total citrus farm fence-mounted PV capacity in Florida is estimated at 12 GW, with a resulting annual energy production of 15 TWh, corresponding to 1,272 kWh/kWp yield (Table 6). The annual yield across the state varies from 1,095 kWh/kW for fixed vertical east-west orientation PV to 1,674 kWh/kW for fixed optimal tilt south oriented PV [63]. These results highlight the importance of fence-mounted PV potential despite the orientation being imposed by existing structures as represented in Figure 5b. Moreover, fence-mounted PV requires no additional land transformation and the results in Table 6 indicate that up to 6% of Florida’s electricity consumption in 2024 (255.1 TWh) could be supplied by citrus farms perimeter PV alone.

If fence-mounted PV concept is extended to all farms in Florida, the total capacity increases almost five folds from the citrus case, reaching a capacity potential of 51.1 GW and 65 TWh of annual energy production as reported in Table 6 and Figure 6. This PV potential can provide 25% of the total annual energy consumption of the state in 2024. The mean PV capacity in this case is 0.8 GW (1 TWh) with a standard deviation of 0.8 GW (1TWh). The large spread in power and energy observed across all farms confirms the impact of farms dispersion on aggregated results reported in citrus farms. Polk County remains the county with the most farms (3,866 post-polygon dissolution) as well as the highest PV capacity potential (3.5 GW) and energy production (4.5 TWh). In contrast, Monroe County exhibits the lowest capacity (0.7 MW) and energy production (0.9 GWh) with only two farms. No farm data was available for Franklin and Alachua counties for the year 2024 as illustrated by the grayed area on the maps in Figure 6.

4. Discussion

Due to the incredible disparity between what FPL is paying for solar electricity (3 cents) and what they are charging customers for it (12 cents), being a grid energy provider does not currently make sense. This underpaying of solar electricity has been discussed in detail in the value of solar literature [64,65]. FPL does build their own PV facilities and provides solar paths for their customers [66], presumably at a profit so there may be some opportunities for FPL to partner directly with groves to do vertical systems. For all the rest of citrus grove owners, to avoid going out of business, they need to form partnerships with large loads. Economics only make sense if it can be behind the meter, otherwise the payback time is beyond the lifetime of the PV under warranty as seen in the results. From a practical standpoint this means that citrus grove owners will need to partner with loads that can be moved on premises to have those energy users pay for the fencing needed to protect the citrus crop. In ideal cases, citrus groves adjacent to large energy consumers like industrial manufacturing would be able to form a joint venture. Combining solar with heat pump technology can allow for other uses as well including food drying, biomass processing, and low-temperature manufacturing [67]. Other groves, not so fortunately located, would need to bring in other options.

Co-locating electrical loads with the grove solar shield generation reduces transmission losses and improves the economics by capturing value behind the meter. A growing body of peer-reviewed research identifies several categories of loads that are particularly well-suited to relocation to solar PV output. First, computing loads are increasingly proposed as solar co-located loads [68,69]. Cryptocurrency mining like bitcoin mining in particular has been studied as a good match for PV systems [70] and can also be mounted in mining farm containers that makes them a deployable load [71]. Conventional data centers with workload-shifting capability (training AI models, batch processing) are similarly attractive for even varying the load [72]. Second, solar groves next to highways [73] or fleet depots can be used for electrical vehicle (EV) charging [74]. Perhaps, most applicable would be using agrivoltaic shields co-located to EV fast-charging hubs [75,76]. Finally, long-haul truck depots and bus fleets with predictable midday dwell times are particularly suitable [77]. Citrus PV systems could also be used for powering controlled-environment agriculture, which uses substantial electricity for LED lighting, HVAC, water pumping and dehumidification [78]. There is already precedent for such coupling of agrivoltaics and indoor vertical farming with the agrivoltaic agrotunnel [79]. In addition the citrus region in Florida is suffering from salt water intrusion [80] and reverse-osmosis desalination and municipal water treatment are energy-intensive but tolerant of variable input power when paired with storage tanks making it a good potential for use of future solar energy generation [81]. Finally, green hydrogen production via electrolysis [82] or coupled with other hydrogen generating systems [83] is a potential flexible load. Hydrogen produced on-site can be stored, transported, or used for fertilizer, fuel, or industrial feedstock. Similarly the PV production could be directly coupled to nitrogen fertilizer production [84,85].

Despite the economic limitations and need for load matching associated with the compensation scheme offered by FPL, the GIS scaling results indicate that vertical agrivoltaic fence systems could transform citrus grove boundaries from passive land edges into dual-use infrastructure that provides both energy generation and ACP mitigation. This is particularly relevant for Florida, where HLB has reduced production, increased maintenance costs, and intensified pressure to abandon or convert citrus land over the past decade. Even under current conditions, using 2023 crop data, fence-mounted PV on citrus groves alone could supply approximately 6% of the state’s electricity demand in 2024. This potential could increase if grove restoration occurs, suggesting a reinforcing feedback mechanism in which protected groves support agricultural recovery, which in turn expands the available perimeter for PV deployment and increases total energy generation. Beyond citrus, the approach can be extended to other crop systems in Florida, where perimeter-based PV could serve as both an energy source and a plant protection strategy. More broadly, these results highlight that farm boundaries, windbreaks, and fence lines represent an underutilized class of linear infrastructure that can support distributed solar deployment, particularly in applications where protection, access control, or biosecurity already justify perimeter investment.

4.1. Limitations

This study only considered the theoretical potential of agrivoltaic shield technologies to protect Florida citrus. The windbreak tested experimentally to date were made of mesh and open to wind flow to avoid turbulence, which would be expected to reduces efficacy [19]. This however needs to be tested. It remains to be determined whether 1200 m of solid 2.27m wall of PV and mesh netting up to the range of flight of ACP is sufficient to protect citrus trees from greening or whether the solid bottom of the shield would cause enough turbulence that could allow the ACP to fly over the wall. Future work is needed with both computational fluid dynamics and experiments to verify the protective effects would be possible without also covering the top surface of the grove with mesh netting as well.

The GIS results should be interpreted as a spatial scaling estimate rather than a deployment-ready design. The analysis assumes full-perimeter coverage, representing an upper-bound scenario; in practice, fence-mounted PV deployments for citrus grove protection may be partial, phased, or targeted to high-risk edges depending on ACP pressure, prevailing wind, access requirements, and farm operations. Continuous PV coverage along mapped boundaries also neglects real-world interruptions such as gates, equipment clearances, setbacks from tree rows, and drainage features. The preprocessing approach, including polygon dissolution and hole removal, avoids double counting of shared boundaries but removes internal edges, which may lead to a conservative estimate of total PV potential in fragmented farm systems. At the modeling level, the use of county-centroid weather data and uniform system assumptions does not capture local variability in irradiance, shading from vegetation, or microclimatic effects. In addition, the representation of vertical PV performance does not explicitly account for soiling, mismatch losses along extended fence runs, or degradation mechanisms specific to agricultural environments, as accurate loss modeling and site-specific performance characterization were beyond the scope of this large-scale GIS assessment but can be integrated into future work

4.2. Future Work

From a system integration perspective, the GIS analysis assumes that PV-based barriers are functionally comparable to existing ACP mitigation strategies, although most prior studies have focused on permeable mesh screens rather than solid or semi-solid structures [86,87]. As a result, the interaction between PV walls, airflow, and psyllid transport remains uncertain and requires experimental validation. Thus, the most critical area of future work is the need for experimental tests to show that this mechanism of protecting citrus groves from greening is technically viable. This means careful experiments in which ACP are not allowed to circumvent the wall. The same double-door or foyer entrances used in CUPS [24] can be adopted for the PV fencing and if experiments determine the necessity because of turbulence then the system can be enclosed by HDPE mesh on the top.

The economic analysis used Florida as an example, but because of FPLs rate schedule and compensation scheme solar is only marginally economic. In other regions, PV generation is compensated at much closer rate to the VOS and should make such shielding systems far more economic. Future work is needed to extend the analysis and example provided here to other citrus regions (e.g. China, Brazil, India, Mexico, Spain, etc.).

Future GIS research can also focus on integrating pest dynamics and airflow modeling to identify priority farm edges where fence mounted PV would be the most effective for limiting pest intrusion. Additionally, further work is needed to couple spatial deployment with techno-economic analysis, extending beyond electricity production to account for the co-benefits of pest mitigation enabled by dual-use fence-mounted PV systems [88].

It should be emphasized that the present techno-economic analysis evaluates only the energy revenues and avoided electricity costs of the agrivoltaic shield system. Two additional economic streams that motivate the shield concept in the first place are not yet quantified: (i) the avoided biosecurity capital expenditure relative to conventional ACP-exclusion infrastructure such as CUPS systems at $43,560/acre [24,27], which the agrivoltaic shield substitutes for through its perimeter physical-barrier function; and (ii) the recovered or preserved citrus yield enabled by reduced ACP transmission of HLB. Even partial recovery of pre-HLB yields would generate annual revenue that substantially exceeds the energy-side cash flows reported here. Once these two streams are integrated into the techno-economic framework, the agrivoltaic shield system is expected to demonstrate substantially stronger economics as an integrated biosecurity-and-energy infrastructure rather than as a pure energy project, positioning the central claim of self-funding biosecurity in the strict sense where the avoided infrastructure cost alone exceeds the system's capital expenditure independent of energy revenue. Quantification of this integrated economic case, drawing on the recent PRISMA review of co-benefits in dual-use solar PV systems [88], is the principal direction of follow-up work.

5. Conclusions

This study evaluated the technical and economic feasibility of vertical agrivoltaic shield walls as a dual-use solution to protect citrus groves from ACP intrusion while generating renewable electricity. Structural analysis confirmed that 2.27 m tall PV fences using A-36 steel posts meet Florida Building Code and ASCE 7-22 wind load requirements at 150 mph design speeds. The Vero Beach case study of a typical citrus grove showed that perimeter-only deployment (Case 1) yields over 600 kW-DC producing 563 MWh/year, while adding interior wind shields (Case 2) increases capacity to nearly 1MW-DC and generation to 861 MWh/year. Energy economic analysis under a three-sub-case capital-cost sensitivity sweep showed that profitability depends jointly on the compensation pathway and the project ownership structure. Under FPL net metering (Scenario A), 25-year bill savings versus the grid-only baseline reach $1.27M (Case 1) and $1.97M (Case 2) across all three CapEx sub-cases, and capital recovery is achieved with substantial margin under the owner-developed deployment: Sub-case 3 yields ROI of 20.34 percent for Case 1, and ROI of 29.51 percent for Case 2. At third-party developer pricing, both cases are below breakeven on energy alone, though Case 2 is close to recovery even under this most-conservative CapEx assumption. Wholesale export under COG-1 (Scenario B) produces substantially worse energy returns across all CapEx sub-cases because the avoided-cost rate is approximately half of the retail energy rate that net-metered exports recover. Citrus growers in Florida must therefore couple agrivoltaic shields with co-located loads matching annual PV output and adopt the owner-developer deployment structure to capture full economic value on energy alone; with both policy levers in place, the agrivoltaic shields are self-financing for both configurations evaluated. GIS scaling analysis showed that statewide deployment on Florida citrus groves alone could provide 11.9 GW of capacity generating 15.4 TWh annually, equivalent to 6% of Florida's 2024 electricity consumption. Agrivoltaic shield walls thus offer citrus growers a self-funding biosecurity infrastructure that can simultaneously combat HLB, diversify farm revenue, and accelerate Florida's clean energy transition. If this approach is extended to all Florida farms over 50 GW and 65 TWh (25.5% of state demand) would be generated while actively improving food output making agrivoltaics fences worth exploring outside of citrus as well.