Sword ferns invade Florida Everglades 
Insert picture here
GIS Applications in Invasive Species Management



By David Brandt — GEO 565 — Fall 2009








Remote sensing image of

 a grove of invasive Albezia in Hawaii

Tamarisk, or Salt Cedar, is a major problem in the SouthwestInvasive Cordgrass in Humboldt county




A Global Problem

The problem of invasive alien species (IASs) is a global one. It appears as more nations develop and begin moving products and people around the globe, more issues (and costs) arise due to IASs causing ecosystem havoc and loss of biodiversity. Fortunately, as more countries are forced to deal with the issue, more people become aware of the problem and want to invest in a solution. As this awareness has grown, so has the need to detect, project, measure, and conceptualize the issue. GIS tools have stepped into this role seamlessly. Research into the role GISs play in the fight against IASs has come from around the globe. It is my hope, through this annotated review of selected literature, to get a grasp of how GISs are being utilized in this fight and what the future may hold for this relatively new application.


 

 

Pino, Joan, et al. “Invasibility of Four Plant Communities in the Llobregat Delta (Catalonia, NE of Spain) in Relation to Their Historical Stability.” Hydrobiologica. 570.1 (2006): 257 - 263.

The authors of this study wanted to find a correlation between changes in the preservation of pristine areas and the presence of invasive alien plant species. This study, done on four separate plant communities within the Llobregat delta in Catalonia, used GIS techniques to extract “naturalness” data from orthophoto maps created in 1956 and 1999. Each photo was vectorized into land cover classes, then combined to create a map of change-in-stability zones. This map of “naturalness” was used to design a field sampling plan within the four plant plots. Vegetation inventories were taken according to community type and amount of stability. Results displayed a relationship between invasibility and stability-as stability decreased, invasibility increased. Reedbeds and fixed dune communities seemed to show this trend the most; indicating that as climax communities developed, alien species were less likely to establish. However, the effect of the relationship was dulled in the halophilic rushbeds and scrub lands, most likely due to the environmentally stressful saline habitat.

Rosso, P.H., et al. “Use of LiDAR to Study Changes Associated With Spartina Invasion in San Francisco Bay Marshes.” Remote Sensing of Environment. 100 (2006): 295 - 306.

This study focused on the use of LiDAR (Light Detection and Ranging) to detect the microtopography created by invasive Spartina sp. The authors hoped to better understand how these changes in topography may be influencing the functioning salt marshes in the San Francisco Bay, as well as be able to foretell and mitigate the changes caused by these aggressive plants. They predicted that changes could be measured within two years of initial invasion. Two tidal salt marshes were chosen for the study and were both flown and ground surveyed on the same day to minimize error for accuracy testing. This study was designed to measure relative accuracies within the LiDAR system. Tests were done on mean elevation and scatter on each surface type at different azimuth angles, and how inaccuracies between return order and different flight lines effected the overall performance on different surface types. Then, a second survey was flown a year later to test for discrepancies between datasets. ArcGIS 8.3 was used to convert LiDAR last-return point data into raster form to create a ground model of possible topography changes. The resulting point data indicated that first and last points had been inverted due to vendor error, but regardless of this, point penetration to ground was extremely low to none for vegetated areas. This suggested LiDAR use in the creation of a “bare earth” model in marsh areas may be compromised due to the extremely low, dense vegetation. However, using variables correlated to vegetation height seem to be a good alternative. Using this method the authors observed a noticeable change in the erosion/sedimentation patterns due to the rapid growth of Spartina sp. into historically clear areas. Expansion was measured to be 2.5m/year which will give future projects a tool in identifying the native species from the invader.

Dillemuth, Forrest P., et al. “Patch Dynamics of a Native Grass in Relation to the Spread of Invasive Smooth Brome (Bromus inermis).” Biological Invasions. 11 (2008): 1381 - 1391.

The influence of IASs on native-plant patch dynamics was the target of this study. With few studies focusing on sub-meter native-invasive patch dynamics, the author devised a six year study using GPS surveying techniques to tease out the change in distribution patterns between invasive Smooth Brome and native Prairie Cordgrass in three prairie fragments near Grand Forks, North Dakota. Each patch was mapped three times between 2000 and 2006. Attention was given to patches of Smooth Brome and Prairie Cordgrass. Layers and subsequent buffers were created in ArcMAP. From these datasets, patch area ratios and rates of change dependent on corresponding species were calculated and finally, extinction/establishment rates for the native cordgrass were measured dependent on the influence of the invasive brome. The authors found a strong relationship between increases in Smooth Brome populations and the weakening of Prairie Cordgrass patch dynamics. Not only were cordgrass patch proportions falling, proportions of the brome were increasing noticeably. However, it is noted that brome has not been shown to completely terminate cordgrass patches, only weaken them. A larger conclusion is made by the author indicating that invasive species, as a whole, are capable of reducing abundances of native species when native patches are invaded and have long lasting effects on lower and higher trophic levels.

Chen, Hao, et al. “Developing Habitat-suitability Maps of Invasive Ragweed (Ambrosia artemisiifolia.L) in China Using GIS and Statistical Methods.” GIS for Health and the Environment: Development in the Asia-Pacific Regions. Berlin: Springer, (2007): 105 - 121.

North American Ragweed is the IAS focused on in this study, within an invaded area of China. The authors proposed to create a prediction model of spatial distribution by producing a habitat-suitability map from the native range and projecting it onto the exotic (invaded) range using spatial analysis and statistical GIS tools. As IASs become more severe in China, an understanding of environment-species relationships becomes more important. An ecological model, made up of topological, climatic, and land cover data (from herbarium specimen records), was designed to the requirements of Ragweed in its native environment. This was done using a binomial logistic regression and an Akaike’s information criterion (AIK) model. From the resulting data a habitat-suitability map was created through a reclass of the data and using ERDAS 8.6. The map was then projected onto the exotic range (China) and predictions show a strong possibility for extensive expansion. Each dataset was shown to have a strong influence on ragweed distribution. The studies results showed that there is a great potential for Ragweed to move quickly to other regions of China not yet effected by the invasion, which displayed how useful this predictive model could be in other IAS situations through international cooperation and data sharing.

Lavoie, Christian, et al. “Geographic Tools for Eradication Programs of Insular Non-native Mammals.” Biological Invasions. 9 (2007): 139 - 148.

The plight of island biodiversity due to IASs, specifically from non-native mammels like goats, was the backdrop for this study. Through the use of many GIS and GPS tools, the authors devised several techniques in which to anticipate, hunt, and monitor feral goats on Santiago Island, Galapagos, as well as provide foundational methodologies for future non-native mammal mitigation. SPOT images, in conjunction with ArcGIS 8.1, was used in the planning, implementing, and blocking of aerial and ground hunts, as well as for the Judas Goat technique wherein a radio-collared goat is set free to lead researchers to remaining colonies and allowing them to actively monitor/anticipate movements. All movements, kills, and quadrants are recorded using mobile GPS units, allowing projection of concentrated and critical areas. The study showed that visualizing and monitoring standard quadrants during hunting and detecting was made more time and cost effective through the use of GIS tools. Additionally, these tools allowed for a higher success rate due to the detection of the last remaining (sometimes hard to find) animals, and for statistics on success and cost to be documented, visualized, and shared easily.

Joshi, Chudamani, et al. “Remote Sensing and GIS Applications for Mapping and Spatial Modeling of Invasive Species.” Weed Technology. 23.1 (2009): 99 - 107.

In this research project, the authors of this study sought to build a summary of different applications of GIS and remote sensing (RS) within in the realm of invasive species management (much like what I am seeking to build.) Throughout the overview, the researchers were interested in what methods were being used to predict distribution of IASs, what kind of sensors and processing were used, whether or not any bias of success toward any category of IASs has developed, how reliable the products are, and can improvements be made. It was suggested that studies are trending toward mapping IASs more locally and with hyperspectral RS scanners with hieghtened pixel size. The authors goes on to suggest that most successful RS was had in the canopy category of species (i.e. top story species of land cover) through canopy dominant species and their interactions with natives and IASs. Less success is being had with understory species and most animal invaders. In regards to predicting potential areas of invasion, the authors found the issue of ecological and biological complexity to be the biggest hurdle. Because so many factors need to be quantified in dealing with habitat and ecological relationships, RS and GIS models need become far more sophisticated before real results are possible. Finally, the author presses that advances in sensor technology and the higher integration of RS and GIS into the wider realm of IAS research will need to happen before progress can be made.

Pande, Archana, et al. “Using Map Algebra to Determine the Mesoscale Distribution of Invasive Plants: the Case of Celastrus orbiculatus in Southern Illinois, USA.” Biological Invasions. 9.4 (2007): 419 - 431.

Using observations from modeling and predicting the presence and absence of Celastrus orbiculatus, the authors of this study hoped to find a good GIS tool for mapping distribution of invasive plants. Landscape characteristics (solar radiation, elevation, distance to roads, etc.) were identified as the basis of invasion dynamics, and they propose the use of “map algebra” (using raster cell values to represent data from numerous other data fields), in the form of raster logistic regression, as a way of accurately processing predictive data on a given IAS of plant. Predicting the presence (88.3%) and absence (85.5%) of C. orbiculatus was relatively successful with elevation and slope representing the highest influencing factors on predictability (but not necessarily directly influencing the species themselves). Through their results, the authors suggest the map algebra model as a highly predictive tool in the management of invasive plants as long as the dynamics of dispersion are understood, the field samples are representative of the species population, and the relationship between species and landscape factors are correctly modeled.

Evangelista, Paul H., et al. “Mapping Invasive Tamarisk (Tamarix): A Comparison of Single-Scene and Time-Series Analyses of Remotely Sensed Data.” Remote Sensing. 1 (2009): 519 - 533.

A comparison of of methods used in detecting Tamarisk from remotely sensed (RS) data is what the authors of this study propose. Tamarisk, being a very hearty invader of the SW United States, can be difficult to accurately detect through RS in the past. However, the researchers consider the use of better sensors and more advanced software models in an attempt to find the best use of the maximum entropy model with RS data derived from several points in the growing season and using both single-scene and time-series analysis. Using Landsat 7 satellite imagery from six times within the growing season, data was collected in both the visible and near-infrared (NIR) spectrums. It was concluded that the time-series analysis using 72 variables was the best method with 96% accuracy in presence/absence predictions. Generally, the best time for collection accuracy was the late season data sets that had less interference from native plants due to dormancy and early season datasets that were aligned with flowering. The authors conclude that with the right corrections in variables, this model, taking advantage of maximum entropy, could be used toward detecting many other IASs.

Underwood, Emma C., et al. “A Comparison of Spatial and Spectral Image Resolution for Mapping Invasive Plants in Coastal California.” Environmental Management. 39.1 (2007): 63 - 83.

The researchers in this study wanted to compare the different detection accuracies and efficiencies of high versus low spectral and spatial resolutions in sensing three invasive plants (Carpobrotus edulis, Cortaderia jubata, and Eucalyptus globulus) in California. Using AVIRIS imagery, four categories were used in sensing areas around Vandenburg Air Force Base: hyperspectral high resolution, hyperspectral low resolution, degraded spectral high resolution, and degraded spectral low resolution. Overall, the most accurate category for detecting was the hyperspectral high resolution at 75%, followed by hyperspectral low resolution at 58%. Detection peak accuracy for the three species was 97%, 82%, and 68% for E. globulus, C. Jubata, and C. Edulis respectively. It was interpreted that species invading areas of low biodiversity could be more accurately detected using lower resolution spatial sensors (30m); compared to areas of high biodiversity where spatial resolution needed to be higher (4m) in order to measure at an acceptable accuracy. Overall though, spectral resolution was found to be of greater influence to accuracy than variance in spatial resolution.

Rew, Lisa J., et al. “Searching for a Needle in a Haystack: Evaluating Survey Methods for Non-indigenous Plant Species.” Biological Invasions. 8 (2006): 523 - 539.

In trying to find the best surveying technique for bio-inventory studies focused on non-indigenous plants, the authors of this study concentrate on seven different ground survey styles to compare and contrast their abilities to detect accurately and efficiently (in terms of time and distance covered) using ArcView 3.2. A 10 X 10 km section of Yellowstone National Park was selected for the study. The seven studies—one biased (termed “seek and destroy”), one unbiased systematic, two unbiased totally random, and three unbiased random stratified along rights of way (RoW.) Analysis showed a significant advantage to the point measurement using systematic (grid) alignment and random alignment in intersecting more patches of non-indigenous plants when considering both random and weight (distance to RoW) distributions. However, in relation to time and resource efficiency, a biased Search and Destroy method of detection may be the best; although it may be the worst in terms of understanding area distribution due to the obvious bias.


 

More Resources

Schoenig, Steve. “California Noxious Weed Control Projects Inventory.” Information Center for the Environment. 2002. Web. http://endeavor.des.ucdavis.edu/weeds/.

This site is basically a link to a clearinghouse for all the who, where, what and how in relation to the different noxious exotic weeds in the state of California (a state that is taking IAS management and research very seriously) and the management surrounding them. As part of the more encompassing National Resource Projects Inventory (NRPI), the California Inter-agency Noxious Weed Coordinating Committee (CINWCC) created this database in order to keep track of the reduction or eradication of noxious weeds within the state and to make information surrounding each project as transparent as possible.

Schnase, John L. ISFS: Invasive Species Forecasting System. Ed. Neal Most. 8 May 2008. Web. http://invasivespecies.gsfc.nasa.gov.

The Invasive Species Forecasting System (ISFS)--made up of the Invasive Species Data Service (ISDS), Invasive Species Analysis and Modeling Service (ISAMS), and the Invasive Species Decision Support Services (ISDSS)--is a joint project between NASA's Office of Earth Science and the USGS. The focus of the ISFS is to provide a foundation for importing personal research data, NBII biological data, NASA and other source remote sensing data, or other source data into the USGS's early detection and monitoring protocols and predictive models to create highly predictive and specific maps detailing IAS distribution and potentially susceptible areas. With this increase in accessibility to appropriate models and framework data, it is the hope of the ISFS creators to provide a base for better, faster, more informed policy decision making and management.

Underwood, Emma, and Ustin, Susan. “CBD Technical Series No.32 Chapter 11.“ Convention on Biological Diversity. Web. 25 Aug. 2008. http://www.cbd.int/ts32/ts32-chap-11.shtml.

An extremely helpful chapter in an extremely extensive technical series by the Convention on Biological Diversity, this source systematically breaks down all the details behind using remote sensing technology to measure IASs. From what type of charateristics are involved in detecting IAS to how accurate they are and how they are pecieved in a GIS, this site is current and informative describing the most recent research into this field of detection. Beyond the chapter, the whole technical series brings into focus the basics of remote sensing and how it can be utilized in many different ecological applications.

Me and tribec
Happy Surveying!