Agrilus biguttatus National Survey Sample Areas for the Conterminous US (Fabricius.) Oak Splendor
Beetle.
Data format: Raster Dataset - ESRI Shapefile
File or table name: a_big_sample_areas
Coordinate system: Albers Conical Equal Area
Theme keywords: : Forest Pest, Forest Insect, Invasive Species, Exotic, Oak Splendor Beetle, Agrilus biguttatus, Susceptibility, sample areas
Abstract: The sample area Theissen were created by intersecting the susceptibility risk potential, by category, with a systematic sampling point grid.
To attain higher levels of certainty in the higher risk categories, sample areas are intensified where risk of susceptibility is highest.
FGDC and ESRI Metadata:
•
•
•
•
•
•
•
Identification Information
Data Quality Information
Spatial Data Organization Information
Spatial Reference Information
Entity and Attribute Information
Distribution Information
Metadata Reference Information
Metadata elements shown with blue text are defined in the Federal Geographic Data Committee's (FGDC) Content Standard for Digital Geospatial Metadata (CSDGM). Elements
shown with green text are defined in the ESRI Profile of the CSDGM. Elements shown with a green asterisk (*) will be automatically updated by ArcCatalog. ArcCatalog adds hints
indicating which FGDC elements are mandatory; these are shown with gray text.
Identification Information:
Citation:
Citation information:
Originators: Downing, M.C.; F.H. Koch, R.A. Haack, F.J.; Sapio, W.D. Smith, D.M. Borchert. 2009. National Risk map products and
documentation for the Oak Splendor Beetle (Agrilus bigutatus). Ft. Collins, CO: U.S. Department of Agriculture, Forest Service, Forest Health
Technology Enterprise Team. http://www.fs.fed.us/foresthealth/technology/invasives_agrilusbiguttatus_riskmaps.shtml
Title:
Agrilus biguttatus National Survey Sample Areas for the Conterminous US.
*File or table name: a_big_sample_areas
*
Publication date: 20100818
*Geospatial data presentation form: ESRI Shapefile
Series information:
Series name: Version 2.0
Issue identification: 20100818
Publication information:
Publication place: Fort Collins, Colorado
Publisher: Marla C. Downing
Online linkage: http://www.fs.fed.us/foresthealth/technology/invasives_agriliusbiguttatus_riskmaps.shtml
Larger work citation:
Citation information:
Originators: Downing, M.C.; F.H. Koch, R.A. Haack, F.J.; Sapio, W.D. Smith, D.M. Borchert. 2009. National Risk map products and
documentation for the Oak Splendor Beetle (Agrilus bigutatus). Ft. Collins, CO: U.S. Department of Agriculture, Forest Service, Forest Health
Technology Enterprise Team. http://www.fs.fed.us/foresthealth/technology/invasives_agrilusbiguttatus_riskmaps.shtml
Description:
Abstract:
The sample area Theissen polygons shown here are created by intersecting the susceptibility risk surface, by category, with a systematic sampling
point grid. To attain higher levels of certainty in the higher risk categories, sample areas
Purpose:
The product’s intended use is to develop a detection strategy for Agrilus biguttatus.
Supplemental information:
This project incorporates methods developed by Coulston, et al. (2006) to develop national scale sampling areas based on the Forest Health
Technology Enterprise Team (FHTET) Agrilus biguttatus Susceptibility Surface. The process involved an equal weighted overlay of the
Introduction potential and the Establishment potential.
Reference
Coulston, John W., Koch, F.H., Smith, W.D., Sapio, F.J. 2006. Developing Survey Grids to Substantiate Freedom from Exotic Pests. Proceedings of
the Eight Annual Inventory and Analysis Symposium. http://www.forestthreats.org/publications/fhm/frankkoch/Developing_survey_grids_to_substantiate_freedom_from_exotic_pests.pdf
Introduction Potential
The Introduction Potential Surface for Agrilus biguttatus was produced for the Conterminous United States (CUS) in 1 square kilometer (km²)
units by the U.S. Forest Service, Forest Health Technology Enterprise Team’s (FHTET) A. biguttatus Steering Committee. The product’s
intended use is to develop a detection strategy for Agrilus biguttatus. Three primary datasets with standardized values from 0 to 10 were
used as variables in the analysis. Each data set (Table 1) was used in a weighted overlay process where Principal Ports = 33.4% and Markets
= 33.3%, and Distribution centers = 33.3%. The final Introduction Potential Surface output values also range from 0 to 10, with 10 having
the highest potential of introduction.
Each of the variables was used to depict potential locations where Agrilus biguttatus could be released into the CUS. To delineate Agrilus
biguttatus potential flight range, a curvilinear distance decay value was assigned with a risk rating of 10 at the source location and decreasing
to 0 at 5 kilometers away (Table 2).
Principal Ports. Source: Army Corps of Engineer, Waterborne Commerce, Foreign Cargo Statistics (1996 to 2003). A summary of imported
tonnage of commodities that use Wood Packing Material (WPM), the packing material associated with Buprestid species interceptions,
recorded in the APHIS Pest Interception Network (PIN) 309 database. Only commodities exported from countries where A. biguttatus is
present were included, countries of origin were not ranked. This point data was converted to 1 km² grid cells. For a list of specific countries
and see commodities see Appendix A.
United States Ports that received Commodities from Countries (listed below) were used:
The Ports shapefiles are the result of querying a data set summarizing 8 years (1996-2003) of foreign marine cargo import information.
These data have been compiled from Army Corps of Engineers waterborne commerce statistics, and then sorted by commodity type, foreign
country of shipment origin, and U.S. port where the shipment arrived.
Markets. Source: Federal Highway Administration, Freight Management and Operations, Freight Analysis Framework, Highway Truck Volume
(HTV) and Capacity Data and Environmental Systems Research Institute’s (ESRI) City polygon Data. Flow/capacity data was used to
determine the number of truck trips occurring within the city polygons, which were then used to define potential markets.
Using a polygon data set from Environmental Systems Research Institute (ESRI) that depicts Cities in the United States an intersection was
conducted. These City polygons were included as standard spatial data with the shipment of ArcGIS ver 9.3 in the year 2005. Next, the ESRI
City Polygons were intersected with HTV. City polygons were selected that received any truck trips.
Distribution Centers. Sources: National Transportation Atlas Database (2003). Distribution centers that handle commodities that likely use
WPM during transport were considered; 1496 distribution center records were used, and 1510 locations were removed from the analysis. The
Distribution Centers Polygons were selected from the ESRI City polygon data set (listed above). Then a distance decay function illustrated in
table 2 was applied to these data. An additional 193 distribution centers were added. Cartesian coordinates were also provided by national
retailers, including: FedEx, IKEA, Kmart, KOHL’s, Lowe’s, OfficeMax, PETCO, Target, The Home Depot, and Wal-Mart.
Analysis
Finally, each data set was used in a weighted overlay process where Principal Ports = 33.4% and Markets = 33.3%, and Distribution
centers = 33.3%.
Table 1
Introduction Variables
Principal Ports
Markets
Distribution Centers
Value Ranges
0 - 10
0 - 10
0 – 10
Table 2
Distance Decay for Probable Flight Range of Agrilus biguttatus
Distance (kilometers)
0 (Source)
> 1 and < = 2
GRID Value
10 (Extreme)
8
(High)
> 2 and < =
> 3 and < =
= > 5
3
4
3 (Moderate)
1 (Low)
0 (Little or No)
Establishment Potential
The Establishment Potential Surface for Agrilus biguttatus was produced for the conterminous United States in 1 square kilometer (km²)
units by the U.S. Forest Service (USFS), Forest Health Technology Enterprise Team’s (FHTET) Agrilus biguttatus Steering Committee; a
multidisciplinary team with participation from USFS, Animal and Plant Health Inspection Service (APHIS), and North Carolina State University
(NCSU).
The Establishment component for A. biguttatus depicts where the pest could survive if it was introduced. If the pest is known to have already
been introduced, it may be desirable to prioritize locations where the pest populations are most able to survive and may be expanding. In
cases where it is unknown whether the pest has been introduced the Establishment Component should be used in conjunction with the
Introduction component to develop a Susceptibility component for A. biguttatus. Supporting biological information was gathered from USDA
Forest Service Research Station experts, scientific literature, and the Exotic Forest Pest (ExFor) website (North American Forest Commission,
2008) http://spfnic.fs.fed.us/exfor/data/pestreports.cfm?pestidval=154&langdisplay=english
Purpose:
The product’s intended use in conjunction with the Introduction Potential Surface is to develop a Susceptibility Potential Surface for A.
biguttatus. Three datasets were used as parameters in the establishment analysis to determine the level of risk (hazard potential) A.
biguttatus poses in areas where it could survive:
1. Natural Host (i.e. Quercus spp from FIA) (Appendix B)
2. Drought (from 2007 – 2009) (Appendix C)
3. Urban Forest
Natural Host
Source: USDA Forest Service, Forest Inventory and Analysis (FIA) program.
Only species of Oaks (Quercus spp) contained in the FIA were considered (Appendix B). The Oak data were used as a presence absence
input. That is, Oak, size, Trees/Acre, or basal area were not considered.
Drought
Source: USDA Forest Service, Forest Health Technology Enterprise Team (FHTET) (Appendix C).
Extreme late spring or early summer drought conditions from the years of 2007 – 2009 were considered. These data were partitioned into 4
classes: 0 = No drought conditions were observed for all three years, 3 = drought conditions occurred for one year in the three year time
period, 6 = drought conditions occurred for two years in the three year time period, and 10 = drought conditions occurred for three years in
the three year time period.
Urban Forest
Source: USDA Forest Service, Forest Health Technology Enterprise Team (FHTET).
The National Land Classification Data (NLCD) types Deciduous Forest or Mixed Forest was used as our urban forest input subset type. These
data were filtered by urban areas as described by the City Light data set (Imhoff et al. 1997). The urban forest subset cell values were
calculated by summing up the total area, in percent, of the NLCD cell (native cell size of NLCD is 30 meters by 30 meters) occupied within a 1
Km2 grid cell. Next, the data were partitioning into ten integer classes (1 – 10) using Jenks’ Natural breaks.
Process
Natural Host was modified by the drought producing a Disturbed Natural Host data set that contains values of 1, 3, 6, and 10. The Urban
Forest was combined with the Disturbed Natural Host via a overly process. If Disturbed Natural Host was spatially coincident with Urban
Forest the cell value was assigned to the Disturbed Natural Host data set.
References
Imhoff, M. L., W. T. Lawrence, C. D. Elvidge, T. Paul, E. Levine, M. V. Privalsky, and V. Brown. 1997. Using Nighttime DMSP/OLS Images of
City Lights to Estimate the Impact of Urban Land Use on Soil Resources in the United States. REMOTE SENS. ENVIRON. 59:105–117.
Susceptibility Potential
The Introduction Potential and the Establishment Potential were combined in an equal-weighted overlay to produce the final Susceptibility
Potential Surface. The final data were partitioned into five susceptibility classes: 1) Little or No, 2) Low, 3) Moderate, 4) High, and 5)
Extreme. For the purposes of sampling the data were combined into four classes from the original five classes: 0) Little/No, 1) Low, 2)
Moderate, 3) High (combination of High and Extreme).
This project incorporates methods developed by Coulston, et al. (2006) to develop national scale sampling areas based on the Forest Health Technology Enterprise Team (FHTET) Agrilus
biguttatus Susceptibility Surface.
National Scale Sample Methods:
1) reclassification of the susceptibility risk potential surface into four classes from the original five classes (0 - Little/No, 1 – Low, 2 – Medium, 3 – High) Table 1;
2) estimating the intensification factor based on the required number of samples and the relative certainty for each risk category; (Table 1)
3) intensifying EMAP’s North American hexagon to develop a systematic point grid for each risk stratum;
4) spatially intersecting the intensified point grids with the corresponding risk stratum;
5) merging each set of selected points from the stratum intersection;
6) creating the sample areas that are semi-regular tessellations of Theissen polygons created from the merged grid intensification points.
For this project, the FHTET Agrilus biguttatus susceptibility risk potential surface was used and reclassed into the required four classes. A total of 1,000 sample areas were used for the
intensification model. Relative certainties were assigned in order to create increasing plot intensifications for increasing risk stratum. A custom Microsoft ©EXCEL application calculates
the intensification factor and sequence number based on the number of points and desired relative certainty. The sequence for the point intensification is determined from a table supplied
by the authors. Table 1 shows the risk class, area by class, and number of sample areas used for this project. The point grid intensification is based on the Environmental Monitoring and
Assessment (EMAP) hexagon for the conterminous United States and is iteratively intensified using a custom ArcView 3.3 application. A new point file is created for each iterative
intensification. The final intensification iterations for each class is intersected with a vector version of the susceptibility surface and merged to create a single point shape file. A sample area
tessellation is then performed from the merged intensification points using a custom function in the ArcView 3.3 application that creates Theissen polygons. These polygons become the
sample areas, where the sample areas are based on the risk class. It is intended that each sample area is given the same number of sample plots.
Citation
Coulston, John W., Koch, F.H., Smith, W.D., Sapio, F.J. 2006. Developing Survey Grids to Substantiate Freedom from Exotic Pests. Proceedings of the Eight Annual Inventory and
Analysis Symposium. http://www.forestthreats.org/publications/fhm/frank-koch/Developing_survey_grids_to_substantiate_freedom_from_exotic_pests.pdf
Table 1. Summary Statistics
Original
Risk
Risk
Class for
Relative
Class
Sampling Certainty
NoData
Little/No (0)
0.1
(0)
Low (1)
Low (1)
0.1
Moderate Moderate
(2)
(2)
0.6
High (3)
&
Extreme
(4)
High (3)
0.7
Total
Area
2
(km )
Sample
Frequency
6,167,233
1,313,944
45
45
263,426
393
45,321
517
7,789,924
1,000
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
*Language of dataset: en
Time period of content:
Time period information:
Single date/time:
Calendar date: 20100309
Currentness reference:
publication date
Status:
Progress: Planned
Maintenance and update frequency: As needed
Spatial domain:
Bounding coordinates:
*West bounding coordinate: -131.718010
*East bounding coordinate: -50.048796
*North bounding coordinate: 54.232833
*South bounding coordinate: 17.231111
Local bounding coordinates:
*Left bounding coordinate: -2356278.5
*Right bounding coordinate: 2257721.5
*Top bounding coordinate: 3172335.3125
*Bottom bounding coordinate: 268335.3125
Place:
Place keywords: Conterminous United States
Place keyword thesaurus: Lower 48 States
Access constraints: None
Use constraints:
None
Point of contact:
Contact information:
Contact organization primary:
Contact person: Marla C. Downing
Contact organization: Forest Health Technology Enterprise Team (FHTET) Forest Health Protection
Contact position: FHTET Lead, Biological Scientist
Contact address:
Address type: mailing and physical address
Address:
2150 Centre Avenue, Bldg A, Suite 331
City: Fort Collins
State or province: Colorado
Postal code: 80526-1891
Country: USA
Contact voice telephone: 970-295-5843
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
Contact electronic mail address: mdowning@fs.fed.us
Hours of service: 9:00 AM - 5:00 PM MT
Data set credit:
Michael F. Tuffly
Steering Committee:
Marla C. Downing, FHTET Lead
Daniel M. Borchert, APHIS PPQ
Frank H. Koch, NCSU
Frank J. Sapio, USFS FHTET
Bill D. Smith, USFS SRS
Robert A. Haack USFS NRS
Roger D. Magarey NCSU
Security information:
Security classification: Unclassified
*Native dataset format: Raster Dataset
*Native data set environment:
Microsoft Windows XP Version 5.1 (Build 2600) Service Pack 2; ESRI ArcCatalog 9.1.0.722
Cross reference:
Citation information:
Originators: Downing, M.C.; F.H. Koch, R.A. Haack, F.J.; Sapio, W.D. Smith, D.M. Borchert. 2009.
National Risk map products and documentation for the Oak Splendor Beetle (Agrilus bigutatus). Ft.
Collins, CO: U.S. Department of Agriculture, Forest Service, Forest Health Technology Enterprise Team.
http://www.fs.fed.us/foresthealth/technology/invasives_agrilusbiguttatus_riskmaps.shtml
Back to Top
Back to Top
Spatial Reference Information:
Horizontal coordinate system definition:
Coordinate system name:
*Projected coordinate system name: NAD_1983_Albers
*Geographic coordinate system name: GCS_North_American_1983
Planar:
Map projection:
*Map projection name: Albers Conical Equal Area
Albers conical equal area:
*Standard parallel: 29.500000
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
*Standard parallel: 45.500000
*Longitude of central meridian: -96.000000
*Latitude of projection origin: 23.000000
*False easting: 0.000000
*False northing: 0.000000
Geodetic model:
*Horizontal datum name: North American Datum of 1983
*Ellipsoid name: Geodetic Reference System 80
*Semi-major axis: 6378137.000000
*Denominator of flattening ratio: 298.257222
Back to Top
Entity and Attribute Information:
Attribute measurement frequency:
As needed
Back to Top
Metadata Reference Information:
*Metadata date: 20100309
*Language of metadata: en
Metadata contact:
Contact information:
Contact organization primary:
Contact person: Marla C. Downing
Contact organization: Forest Health Technology Enterprise Team (FHTET) USDA Forest Service
Contact position: FHTET, Lead and Biological Scientist
Contact address:
Address type: mailing and physical address
Address:
2150 Centre Avenue, Bldg A, Suite 331
City: Fort Collins
State or province: Colorado
Postal code: 80526-1891
Country: USA
Contact voice telephone: 970-295-5843
Contact electronic mail address: mdowning@fs.fed.us
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
Hours of service: 9:00 AM - 5:00 PM MT
*Metadata standard name: FGDC Content Standards for Digital Geospatial Metadata
*Metadata standard version: FGDC-STD-001-1998
*Metadata time convention: local time
Metadata security information:
Metadata security classification: Unclassified
Appendix A
COMM_NAME
All Manufactured Equipment, Machinery and Products
Building Cement & Concrete; Lime; Glass
Forest Products, Lumber, Logs, Woodchips
Primary Iron and Steel Products (Ingots,Bars,Rods,etc.)
Primary Non-Ferrous Metal Products;Fabricated Metal Prods
Sand, Gravel, Stone, Rock, Limestone, Soil, Dredged Material
Paper & Allied Products
Primary Wood Products; Veneer; Plywood
AND
CNTRY_NAME
Algeria
Azerbaijan
Belarus
Czech Republic
Egypt
France
Germany
Hungary
Libya
Morocco
Netherlands
Poland
Russia
Spain
Sudan
Tunisia
Ukraine
United Kingdom
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
Appendix B
FIA Oak Species
FIA Code
801
801
802
803
804
805
806
807
809
810
811
812
812
814
815
803
843
816
817
842
818
819
820
821
822
823
840
824
825
841
826
827
829
844
813
830
831
Scientific Name
Quercus agrifolia
Quercus agrifolia var. oxyadenia
Quercus alba
Quercus arizonica
Quercus bicolor
Quercus chapmanii
Quercus chrysolepis
Quercus coccinea
Quercus douglasii
Quercus ellipsoidalis
Quercus emoryi
Quercus engelmannii
Quercus falcata
Quercus falcata var. falcata
Quercus gambelii
Quercus garryana
Quercus graciliformis
Quercus gravesii
Quercus grisea
Quercus hypoleucoides
Quercus ilicifolia
Quercus imbricaria
Quercus incana
Quercus kelloggii
Quercus laevis
Quercus laurifolia
Quercus lobata
Quercus lyrata
Quercus macrocarpa
Quercus margarettiae
Quercus marilandica
Quercus michauxii
Quercus minima
Quercus muehlenbergii
Quercus nigra
Quercus oblongifolia
Quercus oglethorpensis
Quercus pagoda
Quercus palustris
Quercus phellos
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
845
832
833
834
836
808
835
828
837
838
839
Quercus prinoides
Quercus prinus
Quercus rubra
Quercus rugosa
Quercus shumardii
Quercus similis
Quercus sinuata var. sinuata
Quercus stellata
Quercus texana
Quercus turbinella
Quercus velutina
Quercus virginiana
Quercus wislizeni
Appendix C
Drought Calculation
Agrilus biguttatus: Late Spring-Early Summer Drought 2007-2009
Frank Koch, Bill Smith
We used gridded data (approximately 4 km2 spatial resolution) created with the PRISM climate mapping system to
perform our analyses. The gridded data (WGS72 projection) were downloaded from the PRISM Group web site
(http://www.prism.oregonstate.edu). When these analyses were performed, final versions of total precipitation, mean
daily minimum temperature, and mean daily maximum temperature grids were available for every month from January
1895 until October 2009.
Methods
We adopted an approach, utilizing the PRISM climate grids, in which a moisture index value for a given location (i.e.,
a grid cell) is calculated based on both precipitation and potential evapotranspiration values for that location during the
time period of interest. Potential evapotranspiration measures the loss of soil moisture through plant uptake and
transpiration (Akin 1991). It does not measure actual moisture loss, but rather the loss that would occur under ideal
conditions (i.e., if there was no possible shortage of moisture for plants to transpire) (Akin 1991, Thornthwaite 1948).
The inclusion of both precipitation and potential evapotranspiration provides a fuller accounting of a location’s water
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
balance than precipitation alone. So, to complement the PRISM monthly precipitation grids, we computed monthly
potential evapotranspiration (PET) grids using the Thornthwaite formula (Akin 1991, Thornthwaite 1948):
PETm = 1.6 L(10
[1]
Tm a
)
I
where PETm = the potential evapotranspiration for a given month m in cm; L = a correction factor for the hours of
daylight and number of days in a month for all locations at a particular latitude; Tm = the mean temperature for month m
in degrees C; a = an arbitrary exponent calculated by a = 6.75 ×10-7I3 – 7.71 × 10-5I2 + 1.792 × 10-2I + 0.49239; and I =
1.514
12
T 
an annual heat index, calculated as I = ∑  i 
i =1  5 
, where Ti is the mean temperature for each month i of the year. To
implement Equation 1 spatially, we created a grid of latitude values for determining the L adjustment for any given 4km2 grid cell in the conterminous United States [see Thornthwaite (1948) for a table of L correction factors]. We
calculated the mean monthly temperature grids as the mean of the corresponding PRISM daily minimum and maximum
monthly temperature grids.
We used the precipitation (P) and PET grids to generate baseline moisture index grids for 1910-2009 for the
conterminous United States. Willmott and Feddema (1992) proposed a moisture index, MI′, with the following form:
[2]
P < PET
 P / PET − 1 ,

MI ' = 1 − PET / P ,
P ≥ PET

0
, P = PET = 0

This set of equations yields a dimensionless index scaled between -1 and 1. Though MI′ is typically calculated based on
annual values, for this analysis we were only interested in moisture conditions during late spring-early summer, roughly
a three-month time window of interest. So, we calculated MI′ based on the total P and PET values summed over three
months rather than an entire year. Notably, late spring-early summer represents a different time window depending on
geographic location (i.e., depending on latitude/elevation/climate). For this reason, we actually calculated nationwide
MI′ grids for three different three-month windows during each year 1910-2009: March-May, April-June, and May-July.
(At the end of our analysis, we ultimately combined three output grids for each year into a single grid; details on how
we did this are provided below.)
To determine departure from typical moisture conditions, we first created a normal grid, MI′norm, representing the mean
of the 100 individual MI′ grids generated for each three-month window (i.e., one grid for each year 1910-2009). We
also created a standard deviation grid, MI′SD, calculated from these individual grids as well as the MI′norm grid. We
subsequently calculated moisture difference z-scores, MDZi, from these components:
[3]
MDZ i =
MI 'i − MI ' norm
MI ' SD
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
where i = a particular year in the 100-year period 1910-2009. The MDZi scores may be classified in terms of degree of
moisture deficit or surplus as follows:
MDZi Score Moisture Status
<-2
Extreme drought (2.3 percent frequency)
-2 to -1.5
Severe drought (4.4% frequency)
-1.5 to -1
Moderate drought (9.2% frequency)
-1 to -0.5
Mild drought (15% frequency)
-0.5 to 0.5
Near normal conditions (38.2% frequency)
0.5 to 1
Mild moisture surplus (15% frequency)
1 to 1.5
Moderate moisture surplus (9.2% frequency)
1.5 to 2
Severe moisture surplus (4.4% frequency)
>2
Extreme moisture surplus (2.3% frequency)
To combine the three output MDZ grids for each year (i.e., one each for the March-May, April-June, May-July
windows) into a single nationwide grid, we first subset them using PRISM data related to frost-free period. Briefly, we
divided the conterminous U.S. into three geographic regions (Figure 1) based on the 30-year mean Julian date of the
Agrilius biguttatus National Survey Sample Areas for the Conterminous US
last spring freeze: Zone 1, including all areas with a mean Julian date ≤ 90 (i.e., last freeze prior to April 1); Zone 2, all
areas with a mean Julian date between 90 and 120 (i.e., last freeze between April 1 and April 30); and Zone 3, all areas
with a mean Julian date > 120 (i.e., last freeze after April 30). Next, we matched each three-month window to the most
appropriate zone (Figure 1), and then clipped the corresponding MDZ grid to the zonal boundaries. Finally, we
mosaiced these clipped grids into a single grid covering the conterminous United States.
We re-projected the final output grids to Albers NAD83. For the A. biguttatus risk model (i.e., the map of
establishment risk), we generated binary (0/1) grids for the years 2007-2009 from the final MDZ grids; basically, MDZ
grid cells exhibiting severe or extreme drought (i.e., z-score < -1.5) were assigned a value of 1, while all other cells
were assigned a value of zero. The three binary grids were then added together using map algebra to create a three-level
map of drought risk.