Priyanka Patel NRS 509 12/16/2014 Application of GIS and remote sensing in conservation of vernal pools Vernal pools are small temporary water bodies that contain water for some parts of the year. They lack any permanent inflow or outflow of water. These pools fill up with water during the fall season and may contain water till spring or early summer. The hydro period for these ponds varies from 2-8 months. Due to periodic drying these pools don t support a fish population. This becomes an asset for amphibian population. They use these ponds to lay their eggs. Due to lack of predators the eggs and metamorphs have a better chance at survival than in a permanent pool. In the northeast they are great habitat for wood frog and spotted salamander metamorphs. They lay their eggs in these pools and use the uplands for feeding and forging. Protection of vernal pools has been a problem since they do not fall under the wetlands class. Vernal pools get protection till the edge of the water but development in uplands destroys the amphibian feeding grounds, which forces them to move, and hence they cannot use the vernal pool anymore. Hence vernal pool protection requires providing a buffer around it. The biggest problem with conservation of these pools is the difficulty to find them on the landscape. They are present randomly on the surface like in the middle of forest, shrub lands etc. To carry out field surveys by walking transects in the forest would make mapping for an entire state a very tedious and labor- intensive job. Their small size and short hydroperiod makes its even more challenging to locate them. GIS plays a very important job in helping locate these pools. Various remote sensing and GIS techniques have been used to predict sites of potential vernal pools. Initially flight lines taking vertical aerial photographs were a resource. These different photographs from different flight line were mosaicked together using GIS to create a photomosaic (Holland, 1998). Photomosaic would provide one big picture of the area of interest. The most common practice is the use of interpreters to look at these images and predict presence of small water bodies that could possibly be vernal pools. This method was not very reliable since it could not be reproduced and the results would very depending on the interpreter. It also increased the chance of error of commission and omission (Wu et al, 2014; Lathrop et al, 2005). This method was still used because it gave a good start to start verifying vernal pools and more ponds are found by this method compared to only field surveys. As GIS techniques improve there is a gradual change in the vernal pool prediction process. Better quality orthophotos are used. As the resolution of the orthophotos became better the identification of small pools became easier. Another
technique used to predict sites is by overlaying different layers. Linear regression is done on confirmed vernal pool dataset to know the important factors determining vernal pool locations. These layers are overlaid to analyze areas where the all match the requirement and vernal pool locations are predicted. Different layers like slope, soil, land use, bedrock geology, precipitation etc have been used for this method (Grant, 2005). The most recent technique is the use of LiDAR data. This data is used to predict small depressions on the surface. This can be done using depression analysis. These depressions can be overlaid with different layers to make sure they are not permanent wetlands or on any developed land. This is an effective way to get locations and chances of error of commission and omission are reduced. These points can be confirmed by checking with orthophotos to make sure that these depressions contain water (Wu et al, 2014). These techniques have been used in many different ways to conserve vernal pools. Old aerial photographs have been used to reconstruct the distribution of vernal pools before urbanization. This has been used to analyze the flora, soil type suitable for vernal pools and conservation plans are made based on this information (Bauder and McMillan, 1998). This photo interpretation technique is used to develop the current vernal pool distribution and compare it to the previous distribution and analyze the land use changes. This information can be used to improve urbanization techniques in ways that there is minimal effect on these pools (Holland, 1998). Another method is to develop a database of potential sites that can be used by volunteers. The volunteers can be trained to carry out verification surveys. This will help create a database of confirmed vernal pools that have the short hydro period and are being used by amphibians. These confirmed pools have been provided protection in many different states (Oscarson and Calhoun, 2007). GIS is a very crucial part of conserving vernal pools. Without the currently available GIS techniques it would have been hard to imagine statewide mapping of vernal pools. The future of GIS applications in vernal pool conservation involves improving quality of orthophotos. Even though very good quality has been achieved they are not readily available for every state. Making this data available will help improve the process of mapping the distribution. LiDAR can play a very crucial role in detecting vernal pools. It is very recent and there is not much LiDAR data available. The LiDAR data needs to be updated for every state, which will improve the quality of predictions and reduce the amount of error. GIS and remote sensing have made this process much easier and with the rate of advancement in this field very soon we will be able to predict all the vernal pools in the state without any error.
Annotated Bibliography Qiusheng Wu, Charles Lane and Hongxing Liu. An Effective Method for Detecting Potential Woodland Vernal Pools Using High-Resolution LiDAR Data and Aerial Imagery. Remote Sens. (2014) Volume 6 This paper focuses on trying to find an alternative other than aerial photo interpretation to predict vernal pools. They want to develop a process where the inconsistencies in the results and improve their repeatability. The interesting approach taken by the authors is to use highresolution LiDAR data and combine it with leaf off color-infrared aerial photography to give more accurate predictions. They use Monte Carlo approach to do a depression analysis, which helps distinguish between actual depressions and other objects. The National Hydrography dataset (NHD) was used with a 10-meter buffer around it and any depression falling inside the buffer were eliminated. This removed all the permanent water bodies and eliminated majority pools that could be connected to a permanent inlet or outlet. This data was generated and matched with the existing database of potential vernal pools. This was done to measure the efficiency of the model to predict the location of these pools. The data was also overlaid with the aerial photograph to visually see if there was water. This helped detect the error of commission and omission. This paper made effective use of LiDAR data and reduced the percentage of error in predicting locations of vernal pools. Bradley W. Compton, Kevin Mcgarigal, Samuel A. Cushman, and Lloyd R. Gamble. A Resistant-Kernel Model of Connectivity for Amphibians that Breed in Vernal Pools. Conservation Biology (2007) Volume 21, No. 3, 788 799 This paper brings up a different view for conservation of vernal pools. While most papers aim to conserve the water body and some parts of the surrounding upland, which in turn protects amphibians, this paper considers the importance of connectivity between vernal pools for the metapopulation of salamanders. They stated that since ambystomatids depend on metapopulation to conserve them there is a need to conserve the connectivity between pools. They used kernel estimator to determine local, neighborhood and regional connectivity. They took into account the dispersal distance of the salamander and the resistance caused by the variation in landscape. Based on the results of these three tests individual pools were given a score. The author intended to provide conservation efforts to areas that have biggest group of high score pools. The results showed that the density was more near suburban areas. The author wants to use this information to determine vernal pool hotspots and provide more protection to maintain connectivity between pools. Damon B. Oscarson and Aram J. K. Calhoun. Developing Vernal Pool Conservation Plans at the Local Level Using Citizen-Scientists. Wetlands (2207), Vol. 27, No. 1, 80 95 The density of vernal pools in an area is really high and to verify them would take a long time. Using citizen scientist can help map these pools faster. This paper uses GIS to locate potential sites and get their coordinate to make it easier for citizens to get to the location and verify the site. They used different aerial photos depends on the availability for that area. They used black and white or colored aerial photographs during spring when the leaves were not
present. These photographs were interpreted to develop a data layer with potential site. This layer was overlaid with tax parcel maps to give information of the property owners where the pools where present. This information was important to get permission to survey these pools. The volunteers were trained to verify pools and assign them in three different tiers depending on the amount of amphibian activity. Every town was given access to GIS information of the verified vernal pools and their tier information. Volunteers usually find the sites missed by photointerpretation during field surveys hence reducing error of omission. Towns develop their conservation strategy use the location and tier information, which helps prioritize the conservation needs of various pools. Evan H. Campbell Grant. Correlates Of Vernal Pool Occurrence in the Massachusetts, USA Landscape. The Society of Wetland Scientists (2005) Vol. 25, No. 2 In this paper Campbell discusses the importance of mapping vernal pools to provide them the required protection. Campbell uses surficial geology, soil, land use, topography, hydrology and average annual precipitation data to predict possible locations of vernal pools. GIS is used to collect the different data layers and analyze the possible locations. The model was validated using set of potential points and another set of random points on the landscape. The model was also checked using pre existing data of verified vernal pools. This showed that the accuracy of the model was 64.8%. The analysis showed negative correlation wit developed areas and cropland. Surficial geology showed that pools were located near sand, gravel, fine grained, flood plain and alluvium while bedrock was not a preferred class. Richard G. Lathrop, Paul Montesano, Jason Tesauro, Brian Zarate. Statewide mapping and assessment of vernal pools: A New Jersey case study. Journal of Environmental Management 76 (2005) 230 238 In this paper they use leaf off- springtime color infrared aerial photography from1995 with1m resolution made available by USGS. Leaf off helps locating the vernal pools while spring time is important because most vernal pools will contain water and will be more visible on aerial photographs. 2002 photograph was not used as it was a dry year and many pools would be dry and hard to locate. The interesting part of the paper was that field surveys were done to measure the error of commission and omission. Results showed 88%accuracy and 2% error of commission. The field surveys also showed minimum size that can be mapped is 0.02 ha. Less than this size greatly increases the size of commission and omission. The good part about digitizing vernal pools was that they preferred to make error of commission as these could be later removed after being field verified by trained volunteers. They avoided omitting mapping vernal pools. The areas surrounding the sites were verified to make sure no close by vernal pools are missed out. Robert F. Holland. Great Valley Vernal Pool Distribution, Photorevised 1996. Ecology (1998). This paper focuses on mapping the distribution of vernal pools in several counties California to study the change in distribution. They use vertical oriented true-color 35mm slides that were developed to keep check of utilization of irrigation water. Many different slides are put together to get the view of all the counties required. 70-90 slides were required to cover a 7.5inch quadrangle. The slides used were taken in the early summer making it easier to distinguish
vegetation from vernal pools. ArcGIS was used to digitize each polygon boundary for the project. GIS was used to provide attributes to each pool depending on the amount of disturbance. This map was compared with the historic distribution and obvious patterns were noticed. The basic layout was similar but the pools that were more disturbed were found near the area that has seen most amount of urbanization. Habitat fragmentation is noticed but there is an increase in total amount of acres covered by vernal pools despites the loss of habitat. The assumption is that the previous mapping may not have been extensive due to poor resolution.