Future Development and Water Quality for the Pensacola and Perdido Bay Estuary Program: Applications for Urban Development Planning

Abstract

Land requirements and impacts from future development are a significant concern throughout the world. In Florida (USA), the state’s population increased from 18.8 M to 21.5 M between 2010 and 2020, and is projected to reach 26.6 M by 2040. To accommodate these new residents, 801 km² of wetlands were converted to developed uses between 1996 and 2016. These conversions present a significant threat to Florida’s unique ecosystems and highlight the need to prioritize conservation and water resource protection, both for the natural and human services that wetland and upland landscapes provide. To better understand the relationship between future development and water resources, we used future development and event mean concentration (EMC) models for Escambia and Santa Rosa counties in Florida (USA) to assess impacts from development patterns on water quality/runoff and water resource protection priorities. This study found that if future development densities increased by 30%, reductions of 7713 acres for developed land, 17,768 acre feet of stormwater volume, ~88k lb/yr total nitrogen, and ~15k lb/yr total phosphorus could be achieved. It also found that urban infill, redevelopment, and stormwater management are essential and complementary tools to broader growth management strategies for reducing sprawl while also addressing urban stormwater impacts.

1. Introduction

Land use change is among the most significant factors impacting natural and seminatural landscapes throughout the world today [1,2,3,4]. Pompe et al. projected that even under a moderate estimate of climate/land use change assumptions, flora would be negatively impacted [5]. Human development has been shown to lead to a range of negative wildlife-, water resource-, and climate-related impacts [6,7]. For example, Mallin et al. showed that high percentages of impervious surface were the most highly correlated factor in fecal coliform abundance in watershed pollution [8]. In particular, urban development typically results in increased water runoff and sediment, nutrient, and other pollutant loading in natural water bodies as a result of reduced recharge and increased runoff [9,10,11,12]. In southeast Florida, it has been demonstrated that urban centers have greater runoff than non-urban areas [13].

As human populations expand and the impacts from climate change continue to grow (resulting in changes in human settlement and migration patterns), an understanding of potential future development threats, likelihood, and suitability is essential for conservation and land use planners working to manage future development and maintain water and other natural resources [14,15,16,17]. Specifically, it is critical to understand the relationship between choices about future development and their potential impacts on stormwater runoff and priority water resources. Among the models that may be employed to study these relationships are future development models, stormwater models, and conservation models.

1.1. Future Development Models

Various future development models have been created that attempt to show areas likely to develop based on specific assumptions and criteria. For example, a model used by the Peninsular Florida LCC (Landscape Conservation Cooperative) considered how urban areas were changing, and included conservation priorities from multiple conservation related models and programs, land use data, impacts from climate change and sea level rise, and looked at three different scenarios of land development acreages (30k, 45k, and 57.5k acres per year) [18]. Van Berkel et al. applied the FUTURES model to the South Atlantic States (SAS) to consider urban growth and associated development from 2010 to 2050 [19]. In the “Farms Under Threat” 2.0 report, authors were interested in how population growth and development were impacting the amount of agricultural lands and other natural resources being converted to development [20]. A model that has been used in Florida (USA), and which this project is based on, is the Sea Level 2040/2070 model, which incorporates assumptions about future population growth, sea level rise, and development likelihood or “suitability” to identify a set of future land use scenarios [21].

The Sea Level 2040/2070 model used three scenarios to explore the impacts of future development in the state. A Baseline scenario captured the current state of development, while Trend and Alternative scenarios were used to explore future development and how land use patterns may be affected by decisions about development density, redevelopment/infill, and protection of natural and agricultural resources. A version of the 2040 model was adapted for use in this project as a means of identifying and evaluating alternative future development patterns.

1.2. Pollutant Runoff Models

A frequent consequence of development is the conversion of natural, pervious areas to unnatural, impervious uses. Stormwater volumes are known to increase over impervious surfaces, also increasing pollutant loading. There are several methods for calculating volumes of stormwater runoff and pollutant loading, including the US Environmental Protection Agency’s Storm Water Management Model (SWMM) [22], the event mean concentration equation, HEC-HMS [23], TR-55 [24], and the Rational Method [25] to name a few.

Runoff modeling, with or without pollutant loading estimates, can be accomplished in several ways. These may include models based on peak runoff or continuous flow. They might be hydrologic or hydraulic models, or a combination of the two. They can also include water quality analyses or incorporate best management practice calculations. They could also be simple calculation equations based on established curve numbers.

Equations, such as for event mean concentration, have been used to compare pre- and post-development differences in stormwater runoff volume and pollutant loading estimates. These show a direct and predictable link between urban land use and water quality and stormwater runoff, that have led to the development of low-impact development (LID) strategies within urban landscapes intended to address stormwater volume, velocity, and water quality issues.

1.3. Conservation Prioritization Models

Incorporating conservation modeling into development planning is also valuable when seeking to ensure that not only is development planned in ways that protect the environment of the developed area itself, but also to plan for the protection of habitats surrounding it. Conservation models can include many different datasets depending on the conservation goals. These might include specific animal or plant habitats, water, or other natural resources.

In Florida, detailed models have been developed that identify conservation priorities and resources for use in state and local conservation planning and land acquisition. One of these is the Critical Lands and Waters Identification Project (CLIP) database, which identifies surface water, landscape, and biodiversity priorities for the state [26]. The CLIP model was developed to identify and prioritize lands and waters in Florida that are critical for protecting the state’s natural resources.

1.4. Study Objectives

The goal of this project is to evaluate the relationship between future land use patterns and water resource impacts for the Escambia and Santa Rosa County region in the northwest Florida Panhandle. This included (1) the creation of two future development scenarios, (2) stormwater volume and pollutant loading estimates, and (3) incorporation of water quality and water storage conservation models.

We are unaware of any previous studies that have combined these three components at the regional scale to evaluate future land use plans in a way that also considers stormwater, water quality, and conservation impacts. We feel that a comprehensive study of this nature could be beneficial in informing future land use plans in a more holistic manner. The future development models identified the amount of land converted to residential, commercial, or other developed land use types based on “Trend” and “Alternative” scenarios that incorporated differing assumptions about future density, redevelopment, and conservation. Specifically, it examined the relationship between sprawling and more compact urban development patterns. Stormwater volume and pollutant loading estimates were calculated to demonstrate the potential runoff quantity and total nitrogen (TN) and total phosphorus (TP) quantities resulting from these two scenarios. Water quality and water storage conservation models, developed in a previous project, were incorporated to assess the potential for water resource priority protection as part of future development/land use plans, particularly in the Alternative scenario.

We leveraged existing GIS-based models developed as part of the Sea Level 2040/2070 project [27], existing water runoff models, and recent model results that were developed to identify priority lands for water quality and water storage. Outputs from the project were used by local stakeholders in the region to draft new or revised future land use planning and development policies. These models and the conclusions about their use may also be used in other locations where the protection of natural habitats and water resources needs to be balanced with the management of future land use and population growth.

Stormwater volume and pollutant loading estimates showed an increase from 1% to 4% for the Alternative vs. Trend development scenarios, respectively, over Baseline conditions. An important finding was that comprehensively limiting development in urban areas important for water quality and water storage priorities has the potential to result in more development outside of existing urban landscapes (in lower-density agricultural or natural landscapes). This led to our conclusion that selective urban infill coupled with LID strategies is likely essential as part of a broader suite of land use planning tools for protecting water resources, which should also include land conservation and careful growth management in rural landscapes.

2. Materials and Methods

Throughout this project, we used the best publicly available GIS and tabular data for the analyses that are described here. Within each section, the data inputs are discussed in more detail. ArcGIS Pro v3.2 was used for spatial analysis.

2.1. Study Site

The project took place in what is referred to as the ‘Panhandle’ of Florida (USA). The future development modeling and the stormwater calculations were performed for the two westernmost counties, Escambia and Santa Rosa, and the Water Quality and Water Storage Conservation Priority models were performed in a previous project for Escambia, Santa Rosa, and Okaloosa counties. The city of Pensacola is the most developed in this area, with approximately 53,724 people in 2023 [28]. Compared to other areas of Florida (USA), this area is relatively undeveloped. However, there is rapid development occurring along the Interstate 10 (I-10) corridor, especially in Santa Rosa and Okaloosa counties.

All three counties border the northern Gulf of Mexico and contain the major rivers Perdido, Escambia, Yellow, and Black Rivers; Perdido, Pensacola, and Choctawhatchee Bays, and multiple bayous, lagoons, and sounds. The Escambia, Yellow, Black, and Perdido Rivers are bordered by wetlands. These wetland areas, and the tributaries to these major rivers, are also classified as ‘Flood Zones’ according to the US Federal Emergency Management Agency (FEMA) [29]. Agriculture in the northern portion of the study area includes timber, silviculture, and rangeland. Conservation areas around Eglin Air Force Base consist mostly of upland forests.

Water quality impairments in the study area include fecal bacteria, total nitrogen, total phosphorus, chlorophyll-a, dissolved oxygen, turbidity, and several metals. Fecal bacteria are present in most waterbodies in the urban areas, but also in more inland areas in the four major rivers. Total nitrogen is also present in the upland areas and present, but to a lesser extent, in the developed areas. The total phosphorus and chlorophyll-a impairments are present in all of the bays adjacent to developed areas. Dissolved oxygen impairments are present predominantly in urban areas, but also along the coastal areas of Santa Rosa and Okaloosa counties, and in a handful of spots in the northern portions of the study area. Metals, mostly mercury, are present throughout the rivers and bays in the study area.

2.2. Future Development Model

As noted, an adaptation of the Sea Level 2040/2070 statewide future development model for Florida was used as the basis for land use scenarios and calculation of stormwater impacts within the project study area. This model has been used for planning and public engagement applications in Florida for many years and was developed by the University of Florida GeoPlan Center, 1000 Friends of Florida, and the University of Florida Center for Landscape Conservation Planning, with updates occurring in 2006, 2016, and 2023 [27]. The version of the model used in this study utilized a suitability-based approach with the same methodology employed in the 2023 statewide version of the model. Population projections, sea level rise data for Escambia and Santa Rosa counties, and suitability models were used to identify areas more likely to develop by 2040.

The future development model used in the study relies on several assumptions. The first is that approximately 5 million new residents will move to Florida by 2040, with almost 80,000 new residents moving to the study area based on population projections from the Bureau of Economic and Business Research (BEBR) at the University of Florida [30]. BEBR population estimates were produced using the housing unit method, which utilizes changes in occupied housing units to estimate changes in population. BEBR uses three data sources to estimate the number of occupied housing units in Florida: active residential electric customers, the number of residential building permits from the US Department of Commerce, and the amount of homestead exemptions reported by the Florida Department of Revenue [31]. The second assumption is that sea levels will rise 0.25 m by 2040 using intermediate projections from the National Oceanic and Atmospheric Administration [32]. Areas inundated by a 0.25 m sea level rise were removed from potential future development, and the existing population was expected to relocate. Another basic assumption was that half the affected population would move within the same county or a nearby county (based on proximity of employment and social networks), and the other half were expected to move out of the state [33,34]. The last assumption is that development is more likely to occur in areas with higher suitability values.

2.2.1. Scenarios

Three scenarios, including a Baseline to represent existing land use, were created for the study area to explore the potential impacts of future development, sea level rise, and population growth. The Trend scenario, often referred to as a sprawl or business-as-usual scenario, modeled future development based on existing densities and patterns of development for 2040. The Alternative scenario was created to explore the potential outcomes from more proactive land conservation, increased redevelopment, and higher development densities. See Section 2.2.3. Development Density, below, for further discussion.

2.2.2. Suitability Map

The modeling process begins with data collection and processing, using the best available data for each suitability parameter. Data is collected and/or created for the suitability layers and development masks. The masks remove areas available for potential development, such as already existing urban development, conservation and other easements, and surface waters. Following the creation of masks, a suitability map is created. Suitability measures the fitness of an area of land for a particular purpose [35]. The selection and weighting of the suitability parameters were based on expert opinions developed for the Florida 2060 and Florida 2070 reports. The updated models used in this study included the addition of sea level rise to the suitability criteria. While the Florida 2060/2070 projects considered proximity to the coastline an attractor for future development, the updated models considered areas projected to be inundated by sea level rise, or land in close proximity to them, to be less suitable for development. Individual layers and the final suitability map contain values from 1 to 9, with 9 being the most suitable based on the model assumptions. Multiple factors are used to determine suitability, including proximity to existing development, road density, absence of wetlands, proximity to the coast, preliminary development approvals, proximity to open water, soil types, and proximity to major roads. The suitability map is the result of the weighted combination of individual suitability layers. Lands suitable for development in this project were based on this suitability map, with areas below 3 out of 9 removed from consideration. This assumption was based on the idea that even with increased development pressure, there will be areas less likely to be developed. Future development in the Trend and Alternative scenarios is based on the same suitability map. However, the Alternative scenario excludes future development in priorities 1 to 3 of the Florida Ecological Greenways Network (FEGN) [36], also known as the Florida Wildlife Corridor, as well as Florida Forever Board of Trustees Projects, one of the state’s two primary land acquisition programs, which have been proposed for acquisition based on their opportunity for recreation, outstanding natural resources, or archeological or historical resources (a state designated dataset referred to as the Florida Natural Areas Inventory, or FNAI [37]).

2.2.3. Development Density

Gross Development Densities (GDD) are used to determine the density of development for future population allocations. The American Farmlands Trust’s Future Development Scenarios to 2040 study used two methods to calculate density. For high-density, low to high intensity urban areas and open spaces from the USGS’s National Land Cover Database (NLCD) were used to determine high-density, whereas the low-density calculations relied on US Census block data. Blocks with an average acre per housing unit smaller than the 10th percentile of the farm size distribution for each county were considered low-density development [38]. The FUTURES models developed at North Carolina State University use historical changes in population and the urban footprint of buildings and infrastructure to determine density [19].

This project used a GDD calculated on existing development densities in each county by dividing the number of people currently living in each county by the total number of developed acres. Residential, commercial, industrial, governmental, and institutional parcels were considered developed and identified using the Florida Department of Revenue Use Codes in the 2022 statewide GIS parcel data [39]. The current population was obtained from the Estimates of Population by County and City in Florida, 2023 data produced by BEBR [40]. The density of development for the Alternative scenario was increased by 30% (based on expert opinion in previous studies). The projected population, density of new development, and the suitability map were combined to determine where future development was expected based on each scenario.

2.2.4. Development and Population Allocation

The model determined where new residents were likely to settle by considering two factors: how much land will be needed to support the predicted population growth, and where residents are likely to move to. The model is intended to represent the footprint of all potential future development (residential, commercial, etc.).

The GDD and population projections were used to calculate how much land was needed, and the suitability map determined where development is likely to occur. The Alternative scenario, in addition to an increase in development density, also assumed a percentage of the projected population to be accommodated by redevelopment.

Table 1. Future population projections for Escambia and Santa Rosa counties for 2040.

County	2023 Population Baseline	BEBR (2023) Projection for 2040	Total Population Change	Percent Population Change
ESCAMBIA	33,452	364,200	30,748	9.22%
SANTA ROSA	202,772	251,500	48,728	24.03%

and

Table 2. Gross development densities (people per acre) used for future 2040 population allocation.

County	Trend Scenario	Alternative Scenario (30% Increase)
ESCAMBIA	2.95	3.84
SANTA ROSA	2.11	2.74

provide the population projects and development densities, respectively, used in these calculations.

2.3. EMC Pollutant Runoff Modeling

The event mean concentration (EMC) method of stormwater volume and pollutant-load estimating was used at the two-county scale to estimate the volume of stormwater and the amount of total nitrogen (TN) and total phosphorus (TP) pollutant loading. Using the three raster outputs from the Future Development Modeling (Baseline, Trend, and Alternative), the furthest extent of development models (which came from the Trend scenario) was established, and the total acres of each land use type to this extent in each of the three scenarios were identified.

The composition of the current development footprint, combined with the USDA Hydrologic Soil Group and the existing DORUC land use codes for the two counties, was identified. This resulted in a table of values that indicated how many acres were currently in each of the possible combinations for land use and hydrologic soil group, the variables necessary for using the EMC method for estimating pollutant discharge for the Baseline scenario. Curve numbers were obtained from the Florida Department of Transportation Drainage Design Guide Appendix B [41], and runoff coefficients and concentration values were obtained from the Escambia County LID Manual [42].

Because the development forecasts for the Trend and Alternate scenarios cannot predict what land use will occur where, the percentage of how many acres of new development would fall in each of the land use/hydrologic soil group combinations was calculated using current development values. These percentages were then applied to the total acreage in the Trend and Alternative scenarios to estimate how many acres were in each of the land use/hydrologic soil group combinations for both counties for both of the future-development scenarios. These calculations provided stormwater runoff volume (in acre-feet), and total nitrogen (TN) and total phosphorus (TP) pollution loading (Annual Mass Loading in lb/yr) for both scenarios.

Following the event mean concentration (EMC) method, curve numbers were provided by the Florida Department of Transportation (DOT), and Percent DCIA (Directly Connected Impervious Area) and pollutant values were obtained from the Escambia County LID Manual (2016). Annual Mass Loading (lb/yr) for both TN (total nitrogen) and TP (total phosphorus) were calculated for all three scenarios. The EMC can be calculated as:

$$EMC={\displaystyle \frac{\sum (C_i·Q_i)}{\sum Q_i}}$$

(1)

where

C_i= concentration of the pollutant in sample i in mg/L;

Q_i = flow volume per the event in L;

$\sum (C_i·Q_i)$ = total pollutant mass during the event;

$\sum Q_i$ = total flow volume during the event.

2.4. Water Quality/Water Storage Conservation Priorities Modeling

The Water Quality and Water Storage Conservation Priority Models were developed in a previous project and were completed for the three-county area of the western Florida (USA) Panhandle: Escambia, Santa Rosa, and Okaloosa counties. ArcMap v10.6 was used for spatial analysis.

2.4.1. Water Quality (WQ) Model

A mask of three impaired waters data layers (Basin Management Action Plans (BMAP), Total Maximum Daily Load (TMDL), and Verified Impaired (VerImp)) was created and used to select all National Hydrographic Data (NHD) Flow Lines that intersected within 200 m of the impaired areas. The selected flowlines were themselves buffered by 200 m. The intent was to capture all contributing waters that flowed into the impaired waterbodies and land area within 200 m of those contributing waterbodies. The 200 m threshold is borrowed from the Florida Department of Environmental Protection’s use of 200 m as the distance for which Onsite Sewage Treatment and Disposal System pollution may impact nearby waterbodies [43,44].

The next step was to identify areas that provide potential conservation value based on their existing land cover. The Cooperative Land Cover (CLC) dataset is developed and updated through a partnership between the Florida Fish and Wildlife Commission and the Florida Natural Areas Inventory (Florida State University). The dataset represents the ecological land cover for Florida [45]. The impaired waters mask and the flowlines mask were used to select all land cover types in the CLC data except the following: communications, highway, industrial, industrial cooling ponds, oil and gas fields, rails, sewage-treatment ponds, strip mines, transportation, and utilities. The purpose of this selection was to create the top priority (P1) of land cover types that would be most suitable for conservation.

A similar process was followed to create the second and third priorities (P2 and P3, respectively) of water-quality-significant conservation lands. All land cover within 200 m of the NHD area shapefile was selected, then all land cover within 200 m of all NHD flow lines was added in order to capture other waterbodies that did not intersect with the impairment areas. Subtracted from this were land cover types that included communications, highway, industrial, industrial cooling ponds, oil and gas fields, rails, sewage treatment ponds, strip mines, transportation, and utilities. A new file was created from this selection. From this file, all shapes identical to the shapes in the buffered flowlines were removed (preserving the P1 areas), and a new file was created. Where the value for compatibility in the CLC layer was high, this was recalculated to moderate, and where the value was moderate or low, it was recalculated to low, thus creating the P2 and P3 priorities, respectively. The P1 and P2 + P3 layers were merged, dissolved based on the priority values, and converted to a raster. This is referred to as the ‘infiltration network’ and identifies the top priority as those lands and waters that are adjacent to waters that are already impaired. P2 priorities are those lands that are compatible with conservation and adjacent to unimpaired waterbodies, while P3 are lands less compatible with conservation.

In addition to the infiltration network, the FNAI Significant Surface Waters and Functional Wetlands layers were incorporated. To create the final Water Quality Model, the priority values in the infiltration network were reclassed as P1 to 3, P2 remained 2, and P3 was reclassed to 1. For the FNAI Significant Surface Waters, P1 was reclassed to 3, and all other priorities were reclassed to 1. The FNAI Wetlands P1 was also reclassed to 3, and all other priorities reclassed to 1. The three layers were combined using the cell statistics function with ‘maximum’ as the overlay statistic.

The goal of the Water Quality Model is to identify those lands that are suitable for conservation and have the potential to improve or preserve water quality. Lands suitable for conservation include those that are natural or mostly natural, such as undeveloped lands or non-intensive agricultural lands. Lands with the potential to improve water quality are those located near existing water quality problems and provide a land cover type that could facilitate the ‘cleaning’ of runoff containing pollutants, such as a vegetated land cover that can provide infiltration benefits.

2.4.2. Water Storage (WS) Model

The spatial datasets used in the Water Storage Model include property tax roll data from the Florida Department of Revenue for the three counties in the study area. The “DORUC” (Department of Revenue Use Code) value was used to identify those parcels that were unsuitable for conversion to conservation, such as those already developed (residential, commercial, industrial, etc.). Lands that were indicated as ‘vacant’, agricultural, or parks were considered higher value because they could be used to store water if a significant precipitation event occurred.

Federal Emergency Management Agency (FEMA) Special Flood Hazard Areas are locations having special “flood, mudflow, or flood-related erosion hazards”. These areas were identified as a high priority for potential conservation and could be used as a water storage area.

Two datasets from the Florida Natural Areas Inventory (FNAI), Florida Forever Conservation Needs Assessment, were also used: the Functional Wetlands and the Natural Floodplain. Both were used because these are water areas that, in their natural condition, can provide storage for excess water. Additionally, the state has determined that Functional Wetlands have high value for preservation because they provide access to the Floridan Aquifer, which is the source of most of Florida’s drinking water. Lands that were identified as high priority in the previous layers, and which were also in high priority in these two layers, would have even greater value because of their transmissivity to the aquifer.

Beginning with the county property tax data, parcels that are vacant, agricultural, parks, or in some other way not highly developed, and which could be used for water storage without concern of significant property damage or loss of life, were selected and given a priority value of 9. FEMA floodplain data was used to identify areas that have significant flood hazards and were given the highest priority value of 9. Priority values for both the Functional Wetlands and Natural Floodplain data were reclassed as follows: 1 = 9, 2 = 8, 3 = 7, 4 = 6, 5 = 5, and 6 = 4, to align them with the other layers. All four layers were joined together in an equally weighted manner. The result provided an output with priority values of 8, 7, and 6, which were recalculated to 1, 2, and 3, respectively, for ease of communicating the results.

The goal of the Water Storage Model is to identify those lands that are suitable for conservation and have the potential to provide water storage, such as from precipitation events greater than normal, and which, under current conditions, would result in flooding of the built environment if waters were not stored in an alternate location. Lands suitable for conservation include those that are natural or mostly natural, such as undeveloped lands or non-intensive agricultural lands. Lands with the potential to store water would include naturally existing wetlands, floodplains, or undeveloped areas that could be landscaped to provide storage opportunities in settings more natural than retention or detention basins.

2.5. Incorporating Water Quality/Water Storage Conservation Models

Priority 1 (P1) results for Escambia and Santa Rosa counties from the Water Quality and Water Storage Conservation Models were used to exclude these areas from future development potential in the Trend and Alternative future development models.

2.6. Public Outreach

A separate component of this project was a public outreach and policy development effort. This effort included the development of an advisory group that included representatives from Escambia and Santa Rosa County governments, Pensacola and Milton city governments, the Treasure Coast Regional Planning Council, Healthy Gulf—an NGO focused on environmentally sustainable development in Gulf communities—and the Pensacola and Perdido Bays Estuary Program (PPBEP), the funder for the project. The advisory group was to provide suggestions for revisions to local development plans and regulations that would discourage sprawl and encourage compact development. The advisory group met several times in both in-person and online meetings. Results of the project were shared with the advisory group, which also helped to inform policy suggestions. These suggestions were compiled into a final report, which was then shared with residents during public outreach events. Local media picked up and further shared the outreach events. Additionally, a recording and documentation of the project are available on the 1000 Friends of Florida website (https://1000fof.org/escambia-santarosa2040/).

3. Results

3.1. Future Development Modeling Results

3.1.1. Scenario Population Allocations

Approximately 80,000 new residents are expected to move to Escambia and Santa Rosa counties by 2040 based on medium projections from the Bureau of Economic and Business Research (BEBR) [46]. At current densities, this results in just over 20,000 acres needed to accommodate new population growth. Priority natural and agricultural lands would see a loss of 4000 acres in the Trend scenario. The Alternative scenario, with its emphasis on denser and more compact development, resulted in 12,000 acres needed for development. Protected natural and agricultural lands in the Alternative scenario would increase by 260,000 acres. These priority natural lands are not all expected to be legally protected by 2040; rather, the model illustrates the potential to accommodate future development while avoiding development in priority areas.

Table 3. Acreage and land use comparisons between current (Baseline) development and the Trend and Alternative future development scenarios for Santa Rosa and Escambia counties combined.

ESCAMBIA & SANTA ROSA	2023	% of Total Acreage	Trend 2040	% of Total Acreage	Alternative 2040	% of Total Acreage
Developed	202,067	18.64%	222,390	20.52%	214,677	19.81%
Protected Natural Forest & Silviculture	271,215	25.02%	268,810	24.80%	491,467	45.35%
Protected Other	43,817	4.04%	42,303	3.90%	84,137	7.76%
Natural Forest/Silviculture (Unprotected)	342,434	31.59%	328,697	30.33%	111,406	10.28%
Other (Unprotected)	202,901	18.72%	194,234	17.92%	154,747	14.28%
2019 Open Water	21,403	1.97%	21,403	1.97%	21,403	1.97%
Sea Level Inundation: Protected Lands	0	0.00%	3919	0.36%	4362	0.40%
Sea Level Inundation: All Other Land Uses	0	0.00%	2082	0.19%	1638	0.15%
Total Acreage	1,083,837	100.00%	1,083,837	100.00%	1,083,837	100.00%
Total Land Acreage	1,062,434	98.03%	1,056,434	97.47%	1,056,434	97.47%
Total Sea Level Inundation	0	0.00%	6000	0.55%	6000	0.55%
Total Open Water Including SLR	21,403	1.97%	27,403	2.53%	27,403	2.53%

,

Table 4. Acreage and land use comparisons between current (Baseline) development and the Trend and Alternative future development scenarios for Escambia County.

ESCAMBIA	2023	% of Total Acreage	Trend 2040	% of Total Acreage	Alternative 2040	% of Total Acreage
Developed	109,006	25.47%	113,960	26.63%	111,279	26.00%
Protected Natural Forest & Silviculture	31,082	7.26%	30,666	7.17%	153,975	35.98%
Protected Other	12,521	2.93%	12,211	2.85%	30,380	7.10%
Natural Forest/Silviculture (Unprotected)	174,922	40.88%	172,129	40.22%	50,149	11.72%
Other (Unprotected)	88,993	20.80%	86,118	20.12%	69,301	16.19%
2019 Open Water	11,408	2.67%	11,408	2.67%	11,408	2.67%
Sea Level Inundation: Protected Lands	0	0.00%	726	0.17%	775	0.18%
Sea Level Inundation: All Other Land Uses	0	0.00%	714	0.17%	665	0.16%
Total Acreage	427,932	100.00%	427,932	100.00%	427,932	100.00%
Total Land Acreage	416,524	97.33%	415,084	97.00%	415,084	97.00%
Total Sea Level Inundation	0	0.00%	1440	0.34%	1440	0.34%
Total Open Water Including SLR	11,408	2.67%	12,848	3.00%	12,848	3.00%

and

Table 5. Acreage and land use comparisons between current (Baseline) development and the Trend and Alternative future development scenarios for Santa Rosa County.

SANTA ROSA	2023	% of Total Acreage	Trend 2040	% of Total Acreage	Alternative 2040	% of Total Acreage
Developed	93,061	14.19%	108,430	16.53%	103,398	15.76%
Protected Natural Forest & Silviculture	240,133	36.61%	238,144	36.31%	337,492	51.45%
Protected Other	31,296	4.77%	30,092	4.59%	53,757	8.20%
Natural Forest/Silviculture (Unprotected)	167,512	25.54%	156,568	23.87%	61,257	9.34%
Other (Unprotected)	113,908	17.37%	108,116	16.48%	85,446	13.03%
2019 Open Water	9995	1.52%	9995	1.52%	9995	1.52%
Sea Level Inundation: Protected Lands	0	0.00%	3193	0.49%	3587	0.55%
Sea Level Inundation: All Other Land Uses	0	0.00%	1,368	21.00%	973	15.00%
Total Acreage	655,905	100.00%	655,905	100.00%	655,905	100.00%
Total Land Acreage	645,910	98.48%	645,910	98.48%	645,910	98.48%
Total Sea Level Inundation	0	0.00%	4,560	0.70%	4,560	0.70%
Total Open Water Including SLR	9,995	1.52%	14,555	2.22%	14,555	2.22%

provide the results of the change in acreage for various land cover classes from the Baseline to the Trend and Alternative scenarios. Table 3 provides the calculations for the two-county area, Table 4 for Escambia County only, and Table 5 for Santa Rosa County only. Developed land use increased from the Baseline of 18.64% (202,067 acres) of total acreage to 20.52% (222,390 acres) of total acreage for the Trend scenario, and to 19.81% (214,677 acres) for the Alternative scenario for the two counties. For Escambia County only, this changed from the Baseline of 25.57% (109,005 acres) to 26.63% (113,960 acres for the Trend scenario, and to 26% (111,279 acres) for the Alternative scenario. For Santa Rosa County, the change in developed land use changed from the Baseline of 14.19% (93,061 acres) to 16.53% (108,430 acres) for the Trend scenario and to 15.76% (103,398 acres) for the Alternative scenario. It is worth noting that Santa Rosa County has roughly 1.5 times the amount of total acres that Escambia County does, and is expected to see the greatest amount of population growth between the two counties. In Santa Rosa County, unprotected Natural Forest/Silviculture is estimated to see the greatest change from the Baseline of 25.54% (167,512 acres) to 23.87% (156,568 acres) for the Trend scenario, while protection of this land use type in the Alternative scenario estimates that the protected acres of Natural Forest/Silviculture increase from the Baseline of 36.61% to 51.45%, while unprotected Natural Forest/Silviculture is reduced to 9.34%. For Escambia County, which is smaller and is expected to experience less growth, the estimated change in unprotected Natural Forest/Silviculture from the Baseline of 40.88% to 40.22% for the Trend scenario, and to 11.72% for the Alternative and increasing protected Natural Forest/Silviculture to 35.98% (from the Baseline of 7.26%).

The map panels in Figure 1 display the results of the Trend (a) and Alternative (c) scenarios, with the maps focused on the Pensacola, Pace, and Milton areas for the Trend (b) and Alternative (d) results, respectively. Grey shading represents existing (Baseline) development, and red indicates projected new development for the Trend or Alternative scenarios as appropriate. The dark, medium, and light green areas represent lands from a variety of conservation categories. Dark green represents lands that are already protected within the Florida Managed Areas (FLMA) inventory. The Florida Ecological Greenways Network (FEGN) is a prioritized dataset of public and private functionally connected lands (P1–5, with P1 being the highest priority). This dataset is used for public education and to inform multiple state, regional, and federal acquisition programs [36]. Medium green areas are P1–P3 FEGN priorities and those included in the Florida Forever project list. Light green areas are other potential conservation areas that are unprotected. Blue areas in the maps are where the effects of 0.25 m of sea level rise are anticipated to result in open water. The hatched area that overlays the other layers shows where Natural Forest or Silviculture exists. These results demonstrate that the increasing density parameters used for the Alternative scenario constrain new development for the same population projections much closer to the existing development, resulting in less sprawl and greater preservation of natural areas. In panel (b), new development is more visible outward from Pace and Milton, whereas in panel (d), more area is identified as priority conservation.

Because of the significance of timberlands as a land use type within the two-county study area, a basic overlay was developed between existing working and natural timberlands, and the priority conservation lands layer used in the Alternative scenario. This was conducted to address the point that within the study area, avoiding development impacts in priority conservation areas is important for maintaining Silvicultural and Natural Forest land uses and the economic, conservation, and water resources that these areas provide. Figure 2 shows only the conservation (dark, medium, and light green), Natural Forest/Silviculture (hatched overlay), Baseline development (grey), and predicted open water (blue), without any representation of the new Trend or Alternative development projections to highlight these areas.

Figure 3 is similar to Figure 2. It shows only conservation areas (dark, medium, and light green), with a blue hatch to indicate Priority 1 areas from the Water Quality and Water Storage Conservation Model results. This has been provided to make a similar point that avoiding development impacts in areas important for conservation will also support the protection of water quality and provide water storage services within the two-county region. This will be discussed further in Section 3.3 and in more detail in the Discussion.

3.1.2. Sea Level Rise

According to the intermediate NOAA 2022 projections, Florida will see a rise in sea level of approximately 0.25 m by 2040. This results in a loss of 6000 acres in the study area. Much of this loss is projected to be in currently protected coastal areas, though there are existing developed urban areas that would also be impacted.

Detailed sensitivity modeling was not used in this project; however, variations in modeling results can be seen from the three model scenarios, which introduced changes in GDD, areas not included in developed area estimates, and inclusion/exclusion of redevelopment areas. Furthermore, while validation may be useful, it was not within the scope or budget of this project. This is discussed further in Section 4.

3.2. EMC Modeling Results

The results of the EMC calculations in total and for each of the counties are displayed in

Table 6. Results of EMC calculations for Baseline, Trend, and Alternative development scenarios, included total and percent change (%) of developed acres, acre-feet of runoff, and annual mass loading (lb/yr) of both TN and TP.

					Annual Mass Loading (lb/yr)
	Developed Acres		Acre-Feet of Runoff		TN		TP
	Total	%	Total	%	Total	%	Total	%
Baseline	202,170		477,885		2,368,393		417,033
Escambia	108,848		280,640		1,365,786		242,819
Santa Rosa	93,321		197,244		1,002,607		174,213
Trend	211,553	4%	500,065	4%	2,478,321	4%	436,389	4%
Escambia	113,900		293,666		1,429,179		254,090
Santa Rosa	97,653		206,399		1,049,142		182,299
Alternate	204,074	1%	482,387	1%	2,390,706	1%	420,961	1%
Escambia	109,874		283,284		1,378,653		245,107
Santa Rosa	94,201		199,102		1,012,053		175,855

below, including the percent increase (%) from the current (Baseline) development values, and for both the Trend and Alternative future development scenarios. This information shows that if development were to follow the Alternative scenario, with 30% increased development density, stormwater volume could be reduced by 17,678 acre-feet of runoff (from Trend to Alternative), and TN and TP loads could be reduced by over 87k lb/yr and more than 15k lb/yr, respectively. It is worth noting that generalized calculations for EMC at the regional scale, as performed in this project, may lead to either an over- or under-estimation of pollutant loading estimates.

3.3. Water Quality/Water Storage Conservation Priorities Modeling Results

The results of the Water Quality Model and the Water Storage Model are displayed in Figure 4, panels (a) and (b), respectively. P1—Highest Priority areas are those in red, P2 areas are orange, and P3 areas are yellow.

3.3.1. Water Quality (WQ) Model Results

The Water Quality Model indicates that P1—Highest Priority areas for conservation intended to improve or protect water quality are areas around some of the most significant waterways in the study area (Figure 4a). Additional areas with high conservation priority are those around the bays and barrier islands. P2 priority areas have the least coverage of the three different priority areas and appear around smaller waterways, mostly in rural locations. The lowest priority, P3, covers the greatest extent, though it is almost entirely rural.

The results of the Water Quality Model indicate that the highest-priority areas for conservation are located around major waterways and can benefit water quality by providing significant areas for infiltration of runoff before it reaches open water, thereby providing an opportunity for removal of pollutants before they are deposited into surface waters. These areas are much larger than mere riparian buffers, extending outward from surface waters by several hundred meters. Additionally, the highest priorities extend a significant length of the waterways.

Providing this much area for conservation along and outward from waterways introduces significant opportunities for nitrogen uptake, phosphorus fixation, bacterial breakdown, and the remediation of many other pollutants that could travel across the land surface and into surface waters. Aside from the obvious benefit to water quality is the additional benefit to wildlife for access to water sources and migration pathways. For example, much of the Escambia River corridor, a P1 priority in this model, is part of the Florida Black Bear habitat of interest (Figure 5c). Figure 5 shares various priority ratings from several conservation datasets. Panel (a) presents the P1–P5 priorities for the FEGN (2021) analysis. Panel (b) is the Critical Lands and Waters Identification Project, v 4.0, also showing the P1–P5 priorities. Panel (c), as noted, is the Florida Black Bear Habitat priorities for P1 and P2. Panel (d) is the priority P1–P5 for the Florida Natural Areas Inventory (FNAI). We can see in all of these models that the P1 priorities are similar to the P1 priorities in the Water Quality and Water Storage Conservation Models study, and that these priorities are highest around the major water corridors. This alignment further suggests that conserving land to protect or improve water quality provides multiple complementary benefits.

3.3.2. Water Storage (WS) Model Results

Similar to the Water Quality Model, the Water Storage Model’s highest priority for conservation is areas adjacent to waterways, though with fewer P1 results around the bays and barrier islands (Figure 4b). In this model, P1 covers the greatest extent, followed by P2, which is often adjacent to, or an extension of, P1 in many of the waterways. There are very few P3 priority areas in this model, and they are scattered throughout the study area adjacent to P1 and P2 priorities.

While not as extensive as the P1 areas in the Water Quality Model, the P1 plus P2 areas in the Water Storage Model also extend the length of waterways and extend much further than just riparian buffers, taking advantage of the natural floodplains and wetlands provided for the purpose of storing and absorbing excess precipitation.

Here, too, when we compare the priorities from the Water Storage Model to the other conservation models shown in Figure 4, there is much alignment with the top priority areas for conservation, concentrating especially around the major river networks.

3.4. Incorporated Water Quality/Water Storage Conservation Models

As noted previously, one of the project objectives is to incorporate the Water Quality and Water Storage Conservation Priority Model results to determine the impact of excluding these from development in the Alternative scenario—recognizing the importance of protecting water resource priorities as part of a land use planning process. The result was an unexpected ‘leapfrogging’ of development further from the existing urban centers (Figure 6). This resulted in more expansive greenfield development, since the overall acreage needed to accommodate future population exceeded the available developable area within existing urban areas. This was also partly due to the fact that lands important for water quality and water storage within urban areas were also excluded from development. We will talk later about the fact that this underscores the importance of a suite of water resource protection strategies, including land protection in rural areas and LID in urban areas. Ultimately, because of the intent of the Alternative scenario to accommodate future growth while limiting greenfield development, the areas of “leapfrog development” were excluded from the Alternative scenario, and potential water quality or water storage priority areas were made available for development.

Figure 6 shows how this ‘leapfrog’ effect impacted the Alternative scenario results. In this map, grey represents existing Baseline development, light, medium, and dark green represent different conservation categories, and dark blue represents open water from an estimated 0.25 m of sea level rise. Areas contiguous with existing development and included in the final Alternative scenario are shown in red, and “leapfrog development” that was not included is shown in purple. The calculated area of the “leapfrog” development in purple is 9931 acres.

4. Discussion

There are many examples nationally and internationally that have shown similar results. In Malaysia, Camara et al. [47] showed that 87% of studies reviewed found that urban land uses were a major source of water quality degradation through runoff and erosion. Juma et al. [48] showed that in Kenya, population growth and increases in GDP were associated with increases in pollution in Lake Victoria. In Turkey, hydrologically connected lakes (Mogan Eymir via Ankara) were influenced by development in the capital city [49]. A study of the Kelani River in Sri Lanka attributed high annual averages of BOD, TC, and low DO levels to large discharges of municipal wastewater and urban drainage into river basins [50]. This study also indicated that untreated sewage is a critical input to water pollution, further noting that 40% of the world population is without appropriate sanitation. This can be noted in a study by Campos and Cachola [51], where fecal coliform in surface water and clams was associated with urbanized areas with impervious surface areas of approximately 10.5%.

Within the southeastern US, population increases from 1996 to 2019 ranged from 10 to 49% per state, while some coastal counties increased significantly more, even up to 180%. During the period 1996–2016, land cover in these same areas saw increases in developed land cover, with the greatest losses in undeveloped land cover and also palustrine wetlands [14]. A national study by Freeman et al. [52] linked population growth in coastal counties to changes in sediment, and nutrient and fecal microbial inputs with the potential for changes in streamflow and salinity and algal blooms. Even benthic conditions are influenced by population densities, as was shown by Dauer et al. [53] in a study of Chesapeake Bay. Moreover, Paerl et al. [9] found that coastal watersheds in North Carolina exhibited eutrophication, harmful algal blooms (HABs), and hypoxia, which were tied to greater nutrient inputs and coastal development.

When considering how different development scenarios impact the volume of stormwater runoff and the amount of pollutant loading, we can see from the results that sprawl-type development leads to increases in the volume of stormwater that must be accommodated in other areas. When the development infrastructure includes paved surfaces such as roads, parking lots, driveways, or impervious entertainment and recreation spaces, the increase in imperviousness directs the total area rainwater runoff into concentrated areas (i.e., stormwater basins), reducing the opportunity for infiltration while also transporting the increased pollutants of the urban environment. As the opportunity for infiltration is reduced, it becomes necessary to manage larger volumes of stormwater runoff. Related to this is the increase in the amount of pollutants transported across these impervious surfaces, as the total stormwater volume and pollutant loads are directed into these designated catchment areas. In the Alternative development scenario, the more compact development footprint results in less impervious surface area, less runoff, and more available land for infiltration, reducing and dispersing the introduction of pollutants to the environment.

Whether via sprawl or compact development, communities of all development patterns must grapple with water quality and water storage decisions and management. It has been proven repeatedly that development has negative impacts on water quality due to increased inputs of pollutants, either via higher than natural levels of inputs such as TN and TP, or introduction of pharmaceuticals, or personal-care and cleaning products used in many households and businesses.

The WQ/WS models were originally developed to provide land managers with a prioritized ranking of parcels to choose from when investing limited conservation resources and funding, and in that singular capacity, are helpful as intended. However, incorporating them into future development modeling, as in this project, created an undesirable leapfrog effect, exactly the opposite of the goals of the project. A different and potentially more appropriate approach to the Alternative scenario model could have allowed for the development of water quality or water storage priority areas, but only within existing urbanized areas, with the assumption that these locations would have been developed with LID techniques for managing stormwater impacts.

To realize the benefits of the models as intended, it is recommended that they be used as a supplemental information tool. An example would be to run the future development models alone and work to focus compact development into areas adjacent to or within existing urban areas. Where development might take place in areas identified as high priority per the model results, municipalities might consider creating ‘conservation’ or overlay districts that require development to maintain higher standards of low-impact development (LID) or some other form of ‘environmentally friendly’ development. These might include gravel driveways, native vegetation instead of lawns, allowing for local gardens, including pollinator gardens, rain gardens, bioswales, and additional water catchment strategies. Rainwater capture can be allowed and encouraged for all non-potable needs. This might be achieved through local planning regulations and fee incentives (such as reduced fees) to developers. Fee incentives could have the secondary effect of driving more low-impact development even outside of these districts. These regulations might also limit the amount and construction type of roads, especially roads closest to water bodies, where riparian edges and management are an important consideration. An example of one such community is Windermere, Florida, near Orlando. Windermere has high-value residential properties, and yet much of the town still consists of gravel or dirt roads. Another consideration is to prohibit, if possible, any type of shoreline hardening that prevents the natural interaction of water and land. This allows for critically necessary infiltration options via shoreline vegetation before pollutants have the opportunity to enter surface waters.

To maintain the ongoing benefits of low-impact development, municipalities might consider tax exemptions for homeowners who maintain the pre-development characteristics on at least a portion of their property Consider too that there are two aspects to these applications: one is reducing the input at the beginning, and the second is better ‘processing’ of the input before introduction to the environment. Examples of reducing inputs might include prohibiting the sale of fertilizers during dormant or rainy seasons.

As with all models, there are limitations. The future development models were based on a number of assumptions, such as development density and population growth, that have significant effects on the final model outputs. Additionally, the suitability map is sensitive to the weighting of the factors included in the model. A comprehensive sensitivity modeling approach was not used in this project; however, some changes in parameters were introduced and evaluated in the future development scenarios, including GDD, development exclusion masks, and redevelopment assumptions. Furthermore, while validation may be useful, it was not within the scope or budget of this project, which allowed for only one target point. For this reason, 2040 was chosen for its suitability as a near-term horizon for local planning decisions, as well as consistency with prior statewide modeling efforts.

As part of this, it is also important to note that these models are not intended to predict the location of future development. Rather, they allow us to explore the effects of various land use and policy decisions. For example, to assume that the Trend scenario will occur, as it is presented here, with no additional conservation is likely unrealistic, as is the assumption that all priority conservation lands will be protected, as is presented in the Alternative scenario. In this project, the Trend and Alternative scenarios represent a set of potential futures and are useful for sparking discussion. Future models could benefit from a more detailed effect of parameter values, such as different population estimation methods, including differences in the anticipated local relocation due to sea level rise.

Mentioned earlier was the impact that integrating the WQ/WS model had on the Alternative scenario. Future improvements to these models might include using trend analysis of water quality monitoring results, which could more accurately represent changing conditions, possibly getting ahead of rapidly deteriorating conditions, or recognizing where measures for improvement are showing success. A limitation of both the Water Quality and the Water Storage Models is a lack of soil data, socio-economic factors, and land values. Including soil data in the models could allow for a more informed model result by understanding what hydrologic conditions might or might not contribute to the goals of the models. Including additional parameters such as socio-economic factors or land values could support decision-making by stakeholders or conservation managers when potential projects include funding sources specifically interested in advancing conservation near disadvantaged areas or where the purchase price is a limiting factor.

The future development models demonstrate the potential impacts of growth in Florida and the possibility that increased density and redevelopment may be effective in balancing urban development with the protection of natural and agricultural lands. Future work for the land use models may include studying the impact of redevelopment, development density, and migration data on future development patterns. It would also be useful for urban land planners to understand the impact of stormwater planning and management at the regional scale as future land use planning is conducted. Coupled with water quality goals and flood prevention achieved through the Water Quality and Water Storage Model outputs, future land use policies have the potential to encompass a region-wide, holistic plan for development, stormwater, and conservation. As noted, this includes increasing incentives for and ease of implementation of LID development within urban areas, while also funding and incentivizing conservation of large working and natural priority landscapes outside of city limits, high population areas, or urban service zones. This is an area that could benefit from additional policy research on LID overlay districts and financing mechanisms so that these practices become an integrated part of urban planning and not only a reactive solution to the consequences of ‘business-as-usual’ sprawl-type development.

5. Conclusions

This project is a pilot study that aims to identify the potential water resource impacts from future land use and development scenarios. To highlight this goal, the project models multiple future development scenarios, provides estimates of associated stormwater volume and pollutant loading, and also incorporates a set of water quality and water storage conservation models. Combining these three sets of results into an integrated output was used to evaluate potential outcomes of locating future development. A significant result was that, at least within this study area, if all water resource priorities within existing urbanized areas were excluded from future development, the result would be the potential for “leapfrog” development necessary to accommodate the estimates of acreage needed for future population growth. To support consideration of the compact Alternative scenario and still protect water resources as per the intent of the study, the discussion describes the importance of a suite of integrated water resource and development planning tools and includes suggestions for how decision-makers may incorporate solutions to address the concerns of habitat loss and water quality degradation that results from population growth and development that is not strategically managed. As noted, these strategies may include LID as part of infill and redevelopment efforts to protect urban water resource priorities, coupled with land conservation and management strategies in current rural and natural landscapes. While additional work is needed to further evaluate the policy, planning, and design strategies appropriate for specific water resource priorities and parcels to balance both future development patterns and water resource concerns, we feel that this study has provided an important step toward integrating multiple land use planning efforts with the goal of more holistic land use planning that preserves and conserves key habitats and water resources.