Transcript Integrating disparate datasets - Natural Resource Ecology Laboratory

Synthesizing and Spatially Linking Disparate Datasets
Catherine
1,2
Crosier ,
1Natural
Ecological Science Center, US Geological Survey, Fort Collins, Colorado 80525
Methods
Results
We are creating a relational database using Microsoft Access 2000 to
synthesize data obtained on non-native species from partners
throughout Colorado (Table 1). The NRCS species codes will be
used to standardize species names as different projects include
synonyms for the same species. Then, we will spatially link this
relational database through ArcView 3.2 using a SQL connection
and nested locations of species. For example, a point may be located
within a wildlife refuge which is located within a county. This
connection will allow data to be displayed with maps. These
spatially linked datasets can then be used to create predictive models
for hotspots of invasion and distributions of target species. Finally,
we aim to make our spatial database available on-line using ArcIMS
software to allow users to obtain overall numbers and distributions
of non-natives at any given spatial scale (Figure 1).
We compared county species lists from all 63 Colorado counties,
manager survey results from 14 National Forest Service wilderness
areas, and plot data from 358 multi-scale vegetation plots. These
disparate datasets have different biases associated with them. For
example, the Biota of North America Program (BONAP) dataset,
compiled from herbarium records, recorded 252 non-native plants
in Boulder County while the next largest county was Jefferson
County with 135 species (Figure 2). This higher number may be an
artifact of the location of the state’s herbarium at the University of
Colorado Boulder and of the amount of research being conducted
in the Boulder County vicinity, rather than an actual hotspot of nonnative species within the state.
Objectives
Abstract
Extensive surveys to obtain species distribution data are expensive
and time consuming. In addition, many land managers do not know
what species occur within their management boundaries, let alone
adjacent to them. Yet, there is a wealth of data collected by
government agencies and non-government organizations that, when
synthesized, maximize the use of existing data on species locations
without incurring large costs. We obtained datasets including
species lists, plot data, transect data, point data, and individual
species polygon data that have been collected using various methods
at different spatial scales to meet different objectives. We then
created a Microsoft Access 2000 database that relates these different
datasets and allows us to use them in concert. The database was
subsequently linked to ArcView 3.2 GIS. When combined and
spatially linked, the data improved the completeness of each
individual dataset. Species reported in a specific area by one dataset
were not reported by a second dataset for the same location. The
creation of a synthesized, spatially linked database allows
development of predictive models and maps for the distribution,
abundance, and occurrence of non-native species. These products
can be used by land managers at all levels to aid in the detection,
management, and control of non-native species.
Keywords: Data synthesis, Data synergy, Spatial linking, Non-native
species, Ecological Databases
Introduction
Ecological studies are often conducted over short temporal scales in
specific areas. With synthesis of these smaller studies, the number
and complexity of questions that can be addressed increases. Using
pre-existing datasets also capitalizes on resources already expended,
reducing time and money constraints. Additionally, lists of probable
species, including invasive non-native plants, do not exist for many
public land units (e.g. national wildlife refuges), and distribution
maps are not available for many non-native species. It is impossible
to manage an area without knowing what species may occur there.
In Colorado, approximately 50% of the land is publicly owned.
Many times, agencies and other organizations managing these lands
do not communicate with one another. It is important to know what
species occur on adjacent lands because adjacent lands can be source
populations for invading species.
(1) Synthesize pre-existing datasets, including multiple scale (i.e.,
1-m2 to entire county) information on non-native vascular plant
species in Colorado obtained from individuals, organizations,
and agencies at all levels (i.e., federal to local) in a spatially
linked database.
(2) Use online GIS technology to disseminate information on nonnative species locations to land managers and the public in a
user-friendly and easily understood manner.
Table 1.
Data sets obtained (1-6) and promised (7- 14)
Figure 3: County List Additions
DATA SOURCE
County species list
2. USGS
Multi-scale plot data
3. USFS - Forest Health Monitoring Program
Multi-scale plot data
4. Graduate students
Multi-scale plot data
5. San Luis Valley GIS Authority
Point, line, and polygon data
6. USFS - Wilderness area surveys
Manager survey results
7. Center for the Environmental Management
of Military Lands (CEMML)
Vegetation monitoring transects
8. The Nature Conservancy
Polygon, point, and transect data
9. Colorado Natural Heritage Program
Polygon and point data
10. State of Colorado
Quarter quad data for 20 species
11. CSU Research Scientists
Point data and control data
12. National Park Service- NPSpecies
Park species lists
13. U.S. Fish and Wildlife Service
Refuge manager survey results
14. Bureau of Land Management
• Interactive interface
• Multi-scale analyses
• Spatial analyses
• User friendly
• Multi-scale graphics
• Simplistic
• Outreach tool
Figure 5: On-line Interface
View data and
distribution maps
for single species
or invasion trends
for the whole state
or a subset of your
choice.
Figure 4: Dataset Comparisons
U.S. Forest Service wilderness
areas that identified Cirsium
arvense where the county list did
not.
Boulder county has 252 recorded
non-native plant species, but this
number may be an artifact of the
location of the state’s herbarium
at the University of Colorado
and the amount of research being
conducted in this area.
Web-based Interface
When combining these datasets, a mean of 14 species (range 2 to
46) could be added to each county list from plot data (16 counties
contained plots) for a total of 222 records added to the BONAP
dataset (Figure 3). We analyzed Canada thistle (Cirsium arvense)
as a specific species to compare the datasets and evaluate their
synergy (Figure 4). This non-native species was identified in 4
plots (located in two counties) and 8 wilderness areas (12 counties)
where, simultaneously, the county in which these plots were
located had not identified C. arvense. Some wilderness areas span
multiple counties, so we obtained a likelihood for Canada thistle
presence in those counties. Thus, datasets may be improved when
combined.
Point, polygon, and control data
Figure 2: Dataset Bias
Relational Database
• Normalized database
DATA TYPE
1. Biota of North America Program
Figure 1: On-line Spatial Database
• Hierarchical tables
and Thomas J.
1,2
Stohlgren
Resource Ecology Laboratory, Colorado State University, Fort Collins, Colorado 80523-1499
2Midcontinent
Spatial GIS
Database
Gregory J.
1
Newman ,
Plots that identified Cirsium
arvense where the county list
did not.
Hinsdale county has only 2
recorded non-native plant
species, probably because it has
been less intensively surveyed.
Future Steps
Number of Non-native Plants
29 - 45
2-6
46 - 66
7 - 10
67 - 110
11 - 17
18 - 23
111 - 135
136 - 252
24 - 28
These results indicate that data synergy can be used to improve the
quality of individual datasets. In addition, the synthesized, spatially
linked database can be used to create predictive spatial models for
hotspots of invasion in the state or in a specific land management
unit (Figure 5) and to create trend surfaces for individual species.
Acknowledgements
San Luis Valley GIS/GPS