Florida Solar Power Suitability Map - plaza.ufl.eduplaza.ufl.edu/juna/dcp2002/stud_work/hundersmarck2017paper.pdf · solar panels had to be on a plot of land free of trees, ... Now

Ryan Hundersmarck

DCP 2002: Intro to GIS Final Paper

Florida Solar Power Suitability Map

Abstract

Climate change has inspired many to look for new ways to obtain energy in a sustainable way. This project constrains areas of Florida to the most suitable locations

for solar panel implementation by the Florida state and local governments. Criteria pertaining to solar radiation level, land use, and land type were considered and utilized

to constrain the model, and multiple methodologies were explored. Goal

The goal of the project was to employ ArcGIS to determine and map the most suitable areas for the solar panels in order to provide information to lawmakers and the public and to facilitate political action towards a renewable energy future. Background

As climate change continues to cause devastating effects around the world, it is imperative that the United States shifts towards a clean renewable energy system and eliminate the release of carbon into the atmosphere. Solar energy is one of the fastest growing renewable energy sources in the nation, and is also becoming increasingly cost effective.

Currently, Florida gets an insignificant amount of power from renewable energy sources (3). A majority of the renewable energy is from biomass and hydroelectric (3). Solar energy produced on the utility-scale encompasses less than one-sixth of the state’s large facility renewable energy, and if small-scale customer production is included, they contribute less than one-tenth of the renewable energy (3). Florida expects that by 2024, more than two-thirds of its utility-scale renewable energy facilities will be producing solar energy (3).

Florida’s ample sunlight could provide a substantial amount of energy if we utilized solar technology. In order to proceed with this green energy revolution, it is important to begin construction in the most suitable and cost effective locations. Scope and characteristics of the study area

The scope of this project encompasses the state of Florida. Data pertaining to the entire state of Florida were collected. Florida has a land area of roughly 170,000 square kilometers.

Objectives and Criteria Seven criteria were chosen to constrain the desirable land area for solar panel construction:

1) Areas with average daily solar radiation levels at or above 5.5 kWh/m^2. 2) Areas not located on water bodies. 3) Areas not located on Native American Land 4) Areas not located in state forests. 5) Areas not located on parks or recreational facilities. 6) Areas at least 200m away from animal and wildlife crossings. 7) Areas not located on pineland.

The most important of these criteria is the average amount of solar radiation. A

solar input of 5.5 kwh/m^2/day was deemed the minimum viable amount of daily sunlight. This was an arbitrary amount chosen to exclude adequate sunlight levels and constrain the areas with the most sunlight available. Areas with the most sunlight will maximize energy production and consequently provides the most incentive for investment. In addition to excluding locations below the solar input lower limit, the viable parcels were color-coded by increasing solar input. The parcels with the more solar input per day were marked red, and the parcels with the less input were marked yellow.

Certain areas of Florida had to be excluded due to favorability of land type. The solar panels had to be on a plot of land free of trees, because trees will block the sunlight from reaching the panels. The plot of land also had to be conducive to putting any man-made structure with ease. Marshlands for instance, are unacceptable because there are accessibility issues as well as flooding concerns. Grasslands and pasture ranges are perfectly suited for solar panel implementation because of their lack of trees, sturdy land, and their lack of interference with biodiversity. Beach land is highly beneficial because it is public land that receives a significant amount of sunlight. The solar panel farm cannot be located on a water body. The water bodies layer was therefore removed from the Florida map. Native American land was designated off limits because unless the solar energy is for the benefit of the Native Americans, it would be yet another example of the United States government utilizing Native American land for resource extraction.

Parcels that were located near Wildlife crossings were excluded due to the harm that structures pose to the native wildlife. Harmful structures include fencing, monitoring facilities and the solar panels themselves. The considerations also involved distance from cities and roads. Cities often require the more energy for a given land area compared to rural areas, and therefore must be prioritized. A close proximity to major roads will allow for the most cost-effective transportation of supplies.

MethodologyFigure1

ProcessFlowDiagram

Data sets were retrieved from the Florida Geographic Data Library website. The

data included the mapping of state forests, pinelands, wildlife crossings, water bodies, Native American land, and parks/recreational facilities.

A list of zip codes with corresponding values for average solar radiation per day was obtained for every state in the US. The zip codes for Florida were extracted out and

converted into a table that was then uploaded to ArcMap. The table had two fields: zip codes and average solar radiation per day. As shown in Figure 1, the table was joined with a shapefile of Florida zip codes. This is layer was named the “SolarRadiation” Layer. The units for the average amount of solar radiation were kwh/m^2/day.

The parcels with an average solar radiation input of at least 5.5 kWh/m^2/day were selected using the ‘Select by Attributes’ function. The Solar Radiation layer was chosen from the drop-down menu. The method was set to ‘create new selection’. The field labeled ‘Avg_kWh_m2’ was chosen. The greater than or equal to sign, ‘>=’, was chosen. Finally, ‘5.5’ was typed after the greater than or equals sign. The operation was verified and applied. The new selection was converted into a layer using the “create layer from selected feature” tool, found by right clicking the layer and hovering the cursor over ‘selection’. The new layer was labeled ‘Solar5.5’, which represents all parcels with solar radiation levels at or above 5.5 kWh/m^2/day. The land area that possesses favorable solar inputs encompasses almost 37,400 square kilometers of land. The resulting map is shown in Figure 2 below (A map of Florida, Figure 2a, is included for comparison): Figure2a Figure2

MapofFlorida Solar5.5(>=5.5kWh/m2/day)

ThelayerwithdataonthewaterbodiesinFlorida,‘WaterBodies’,(Figure3a)

waserasedfromthe‘Solar5.5’layer(Figure2).Thenewlayerwaslabeled‘Solar5.5_Water’andisdisplayedinFigure3.Thetotallandareacoveredbyfavorablesolarinputswhichisnotlocatedonawaterbodyisalmost28,400km^2.

Figure3a Figure3

WaterBodies Solar5.5_Water

TheNativeAmericanlandareaswereerasedfromthislayer

(Solar5.5_Water),andthenewlayerwasdenotedas‘Solar5.5_Water_NAL’(Figure4).Theresultinglandareawasjustover28,200km^2.

Figure4a Figure4

NativeAmericanLand Solar5.5_Water_NAL

Thestateforestlayer,‘StateForests’waserasedfromthe

‘Solar5.5_Water_NAL’layer.Thenewlayerwasnamed‘Solar5.5_Water_NAL_SF’.Thenewlandthatnowmeetsthecriteriaofexcludingstateforestshasanareaofapproximately27,800km^2.

Figure5a Figure5

StateForests Solar5.5_Water_NAL_SF

The parks and recreation layer (Figure 6a) was erased from the previous layer,

‘Solar5.5_Water_NAL_SF’ (Figure 5). The new layer was named ‘Solar5.5_Water_NAL_SF_Parks’ (Figure 6).

Figure6a Figure6

ParksandRecreationalFacilities Solar5.5_Water_NAL_SF_Parks

Abufferwasplacedaroundthewildlifecrossingusingthebuffertoolinthe

Arctoolbox.Theresultinglayerwaserasedfrom‘Solar5.5_Water_NAL_SF_Parks_WL’.

Figure7a Figure7

WildlifeCrossings(with200mBuffer) Solar5.5_Water_NAL_SF_Parks_WL

Pinelands data needs to be erased from the current layer. The Pinelands layer was

erased from ‘Solar5.5_Water_NAL_SF_Parks_WL’. The new layer was labeled Solar5.5_Water_NAL_SF_Parks_WL_Pine’. All the criteria has now been met and is displayed in Figure:

Figure8a Figure8

Pinelands AllCriteriaMet

The layer was converted to raster using the ‘Polygon to Raster’. The ‘Solar Radiation 5.5’ layer was chosen for the input features. The value field chosen was ‘Avg_kWh_m2’. The cell assignment type was designated as ‘CELL_CENTER’. The cell size was set at 2900. The layer was labeled ‘Solar Radiation_PolygontoRaster’. Figure9

SuitabilityModel

Change in Methodology

Now that the suitability model is completed, an additional methodology was implemented where proximity to major roads and urban areas are desirable. The new process flow diagram is shown in Figure 10.

Figure10

ProcessFlowDiagram

DataforthemajorroadsandurbanareaswereuploadedtoGIS.Thetwo

layerswerebufferedandnamed‘UrbanBuffer’and‘RoadBuffer’.Theurbanareaswerebufferedby3kmandthemajorroadswerebufferedby1km(Figure11and12below).

Figure11 Figure12

Urban3kmBuffer Roads1kmBuffer

Eachbufferlayerwasintersectedwiththeoldsuitabilitymodellayer.(Figure

8).

Figure13 Figure14

UrbanIntersectwithoriginalmodel RoadsIntersectwithoriginalmodel

ThetwointersectionswerecombinedusingtheUnionfunctioninArctoolbox.Nowthemodelisconstrainedtoplacesthatarelocatedeitherwithin1kmofaroadorwithin3kmofanurbanarea(Figure13and14below).

Figure15

Union

Similartothepreviousmodel,thesuitabilitymodelwasconvertedfrompolygontoraster.Thesamevaluefieldandcellassignmentwereused.TheresultinglayerisshowninFigure16below.Figure16

NewSuitabilityModel

AnautomatedmodelwascreatedinArcGISinordertofacilitatethealteringofthecurrentmethodologyandusemorecriteriatoachieveadifferentgoal.ThemodelisshowninFigure17below.TheautomationbeginsafterSolar5.5isalreadycreated,soasubsequentprojectwouldrequirethecreationofitsowntableofzipcodesandaverageradiationlevels.ThelinkprovidedinthereferencespossessesthelinktoalistofsolarradiationforthezipcodesallaroundtheUS.ThezipcodesforFloridawouldhavetobeextractedoutandplacedintoatableonArcGIS.Fromthere,thetablewouldhavetobejoinedwithaFloridaBoundarylayerwhichcanbefoundattheFGDLwebsiteprovidedinthereferences. Theautomationbeginswithmultiplesubtractionsfromthelayerwithatleast5.5kWh/m^2/day,Solar5.5,usingtheerasetool.Theparameterserasedarethecriterialistedearlierinthepaper.AbufferwascreatedfortheWildlifeCrossings.Theresultinglayerwithallthecriteriawasconvertedtorasterformat.Thereasonforthisistomarkthehigherradiationlevelswithdarkercolorstobeeasilyunderstood. Thenewsuitabilitymodelcanbeseenonthebottomrowoftheprocess.Itincludestwoadditions,urbanareasandmajorroads,eachwithabuffer.Thenewsuitabilitymodelwasconvertedtorastersimilartotheoriginalsuitabilitymodel.

Discussion

Unfortunately, I learned that much of the criteria used to constrain the model were proven to be largely inconsequential. Criteria such as Native American Land and wildlife crossings were insignificant and did not constrain the model to any noticeable amount. The Native American land excluded 161 square kilometers out of over 28,000 km^2 of land. The wildlife crossings did not exclude any land area.

Limitations The most glaring limitation of this project is lack of land use data. There are a few criteria that involve land type – pinelands, state forests and water bodies. However, every type of land should be taken into account to develop the best model. Additionally, there might be some places pinelands or state forests that have clearings suitable for solar panels, information not available in the FGDL data. Information about governmental subsidies and local solar panel laws could be helpful for businesses looking to invest in solar energy.

Parks and recreational facilities could have been designated as favorable locations, since solar panels could be placed on the roofs of park pavilions and recreational facilities. Park pavilions and other structures would be a good place to put solar panels because it provides both shade and energy to the park-goers. Park-goers could use power hook-ups for electric grills, phone chargers, and potentially electric car hook-ups in the parking lot. The problem is that not all parks are the same, some parks could be water parks or other facilities of which only a small portion of the land area could be sustain solar panels. The data from FGDL does not specify this information.

Parcels with at least a minimum land area could have been prioritized. This would have identified the parcels best suited for an entire solar farm. This would be conducive to a higher energy output and would make a potential project more desirable. This could have been achieved using select by attributes.

Public facilities could have been prioritized - including school buildings. School roofs I imagine would be suitable for solar panels because the schools could use the energy directly. The idea behind this methodology is that it is easier to convert what the government has complete control over, to minimize red tape.

Conclusion

In conclusion, the need for a change towards renewable solar power is evident in the scientific consensus of the impact of climate change. The barrier to implementation is often political, which is why this project was geared towards illuminating suitable areas of Florida that can begin solar implementation with the best return on investment.

References 1) List of Radiation by Zip Code can be found here (Under Solar Summary near the bottom of the page) http://www.nrel.gov/gis/data_solar.html 2) Data was obtained from FGDL: http://www.fgdl.org/metadataexplorer/explorer.jsp 3) https://www.eia.gov/state/analysis.php?sid=FL