Search

Search Results

• 1. [Article] Landsat-based monitoring of annual wetland change in the main-stem Willamette River floodplain of Oregon, USA from 1972 to 2012

  Despite holding substantial ecological value, wetlands in the United States have experienced a significant decline in both area and function over the past century with the majority of freshwater wetland ...

  Citation × Citation Citation Abstract Copy Title: Landsat-based monitoring of annual wetland change in the main-stem Willamette River floodplain of Oregon, USA from 1972 to 2012 Author: Fickas, Kate, Colleen Despite holding substantial ecological value, wetlands in the United States have experienced a significant decline in both area and function over the past century with the majority of freshwater wetland loss attributed to agricultural conversion. Agriculture is the second largest industry in the State of Oregon and the State places substantial emphasis in its land use planning goals on the preservation of agricultural land. Oregon’s Willamette Valley accounts for the majority of agricultural output with 53% of the valley bottom classified as agricultural land. Additionally, the valley houses 70% of the state's population. The valley was once comprised of extensive wet prairies and abundant riparian forests along the Willamette River floodplain, but native ecosystems have been reduced to a fraction of their original distribution since Euro-American settlement in the mid 1800s. The few wetlands that remain are at high risk to loss and degradation from agricultural activity. Following national wetland conservation policies, Oregon has since attempted to monitor and regulate losses due to disturbance and modification of the State's remaining wetlands through a "no-net-loss" policy aiming to decrease wetland losses and replace disturbed wetlands through mitigation. The National Wetlands Inventory (NWI) was designed to produce detailed maps and status reports of the characteristics and extent of the nation's wetlands and help determine the efficacy of no-net-loss policy implementation on the nation’s wetlands. In some cases, the NWI has been found to have low categorical and spatial accuracy and coarse temporal resolution, with some maps over two decades old. Although Landsat satellite imagery was originally found to lack the needed spatial resolution for classification detail and wetness designation that aerial photography provided, Landsat has 40 years of freely available, high quality annual imagery and should be explored for use in annual wetland change detection. Our objectives were to: (1) Quantify and characterize spatial and ecological trends in annual wetland change through gain, loss, and conversion in the Willamette Valley; (2) Evaluate the effect of the no-net-loss federal wetland conservation policy change enacted in 1990 on trends in net wetland area; and (3) Describe a new methodology that reaches back through the over 40-year Landsat archive to map fine scale wetland and related land-use changes from 1972-2012. We used annual Landsat MSS and TM/ETM+ images from 1972 to 2012 to manually interpret loss, gain, and type conversion of wetland area in the two-year inundation floodplain of the Main-Stem Willamette River using TimeSync, Google Earth, and ArcMap. By creating Tasseled Cap Brightness, Greenness, and Wetness indices for MSS data that visually match TM/ETM+ Tasseled Cap images, we were able to construct a complete and consistent annual time series and utilize the entire Landsat archive. Additionally, with an extended time series, we were able to compare trends of annual net change in wetland area before and after the no-net-loss policy established under Section 404 of the Clean Water Act in 1990. We found that wetlands experienced annual loss, gain, and type conversion across the entire study period. Vegetated wetlands (emergent and riparian wetlands) experienced a 314 ha net loss of wetland area across the 40 year study period whereas non-vegetated wetlands (lacustrine and riverine wetlands) experienced a 393 ha net gain. All wetland types combined saw a 79 ha net increase in wetland area across the full study period. The majority of both gain and loss in the study area was attributed to and from agricultural conversion followed by urban land use. Time series analysis of the rate of change of net wetland area was calculated using the Theil-Sen (TS) Slope estimate analysis. For annual change of wetland area before and after 1990 no-net-loss policy implementations, the rate of annual wetland area lost slowed for riparian wetlands and reversed into trends of annual net gain in area of emergent wetlands. The rate of annual net area gained for lacustrine wetlands was slowed post-policy. Accuracy assessment of land use change polygons in the field was only able to capture 12% of our interpretations due to access restrictions associated with private land. In spite of a low sample size (n=45), overall accuracy of land use classification through wetland change polygons was at 80%. This accuracy increased to 91.1% when land use classes were aggregated to either wetland or upland categories, indicating that our methodology was more accurate at distinguishing between general upland and wetland than finer categorical classes. Title: Landsat-based monitoring of annual wetland change in the main-stem Willamette River floodplain of Oregon, USA from 1972 to 2012 Author: Fickas, Kate, Colleen Despite holding substantial ecological value, wetlands in the United States have experienced a significant decline in both area and function over the past century with the majority of freshwater wetland loss attributed to agricultural conversion. Agriculture is the second largest industry in the State of Oregon and the State places substantial emphasis in its land use planning goals on the preservation of agricultural land. Oregon’s Willamette Valley accounts for the majority of agricultural output with 53% of the valley bottom classified as agricultural land. Additionally, the valley houses 70% of the state's population. The valley was once comprised of extensive wet prairies and abundant riparian forests along the Willamette River floodplain, but native ecosystems have been reduced to a fraction of their original distribution since Euro-American settlement in the mid 1800s. The few wetlands that remain are at high risk to loss and degradation from agricultural activity. Following national wetland conservation policies, Oregon has since attempted to monitor and regulate losses due to disturbance and modification of the State's remaining wetlands through a "no-net-loss" policy aiming to decrease wetland losses and replace disturbed wetlands through mitigation. The National Wetlands Inventory (NWI) was designed to produce detailed maps and status reports of the characteristics and extent of the nation's wetlands and help determine the efficacy of no-net-loss policy implementation on the nation’s wetlands. In some cases, the NWI has been found to have low categorical and spatial accuracy and coarse temporal resolution, with some maps over two decades old. Although Landsat satellite imagery was originally found to lack the needed spatial resolution for classification detail and wetness designation that aerial photography provided, Landsat has 40 years of freely available, high quality annual imagery and should be explored for use in annual wetland change detection. Our objectives were to: (1) Quantify and characterize spatial and ecological trends in annual wetland change through gain, loss, and conversion in the Willamette Valley; (2) Evaluate the effect of the no-net-loss federal wetland conservation policy change enacted in 1990 on trends in net wetland area; and (3) Describe a new methodology that reaches back through the over 40-year Landsat archive to map fine scale wetland and related land-use changes from 1972-2012. We used annual Landsat MSS and TM/ETM+ images from 1972 to 2012 to manually interpret loss, gain, and type conversion of wetland area in the two-year inundation floodplain of the Main-Stem Willamette River using TimeSync, Google Earth, and ArcMap. By creating Tasseled Cap Brightness, Greenness, and Wetness indices for MSS data that visually match TM/ETM+ Tasseled Cap images, we were able to construct a complete and consistent annual time series and utilize the entire Landsat archive. Additionally, with an extended time series, we were able to compare trends of annual net change in wetland area before and after the no-net-loss policy established under Section 404 of the Clean Water Act in 1990. We found that wetlands experienced annual loss, gain, and type conversion across the entire study period. Vegetated wetlands (emergent and riparian wetlands) experienced a 314 ha net loss of wetland area across the 40 year study period whereas non-vegetated wetlands (lacustrine and riverine wetlands) experienced a 393 ha net gain. All wetland types combined saw a 79 ha net increase in wetland area across the full study period. The majority of both gain and loss in the study area was attributed to and from agricultural conversion followed by urban land use. Time series analysis of the rate of change of net wetland area was calculated using the Theil-Sen (TS) Slope estimate analysis. For annual change of wetland area before and after 1990 no-net-loss policy implementations, the rate of annual wetland area lost slowed for riparian wetlands and reversed into trends of annual net gain in area of emergent wetlands. The rate of annual net area gained for lacustrine wetlands was slowed post-policy. Accuracy assessment of land use change polygons in the field was only able to capture 12% of our interpretations due to access restrictions associated with private land. In spite of a low sample size (n=45), overall accuracy of land use classification through wetland change polygons was at 80%. This accuracy increased to 91.1% when land use classes were aggregated to either wetland or upland categories, indicating that our methodology was more accurate at distinguishing between general upland and wetland than finer categorical classes. Close

• 2. [Article] Multi-spatial scale representation of landscape transitions using landsat thematic mapper data and scale-space filters

  This thesis considered current approaches to describing landscape pattern, identified scale issues associated with defining objects, and explored techniques to reliably group elements based on land cover ...

  Citation × Citation Citation Abstract Copy Title: Multi-spatial scale representation of landscape transitions using landsat thematic mapper data and scale-space filters Author: Lindeman, Dale R. This thesis considered current approaches to describing landscape pattern, identified scale issues associated with defining objects, and explored techniques to reliably group elements based on land cover as represented by satellite imagery. It was recognized that there is an important need to develop tools that can be applied using remotely sensed data to objectively identify landscape features at multiple spatial scales. The thesis focused on the potential for the use of specific digital image processing techniques to objectively generate multi-scale landscape heterogeneity models from Landsat Thematic Mapper (TM) satellite imagery. An approach using Difference-of-Gaussian (DoG) scale-space filters was used to process TM satellite data, transformed into the Normalized Difference Vegetation Index (NDVI), to define discrete landscape units across a range of scales. A series of Difference-of-Gaussian filters, with standard deviation (a) pairs of 0.9 and 1.1, 0.8 and 1.2, and 0.7 and 1.3, was applied to TM data. Results were compared against published ecoregion classifications and aerial orthophotography. These comparisons showed that the applied filters failed to accurately define landscape elements at a range of spatial scales. Filter size was found to be highly sensitive to the resolution of the image and object boundaries were maldelineated when complex boundary combinations were encountered. A modified approach using Gaussian/Difference-of-Gaussian or GaussianlLaplacian filter combinations and an alternate approach of watershed segmentation were recommended as potential techniques to test in further study. Future study should also include statistical approaches to aggregate zones and relate delineated zones to an actual, on-the-ground physiognomic classification scheme. Title: Multi-spatial scale representation of landscape transitions using landsat thematic mapper data and scale-space filters Author: Lindeman, Dale R. This thesis considered current approaches to describing landscape pattern, identified scale issues associated with defining objects, and explored techniques to reliably group elements based on land cover as represented by satellite imagery. It was recognized that there is an important need to develop tools that can be applied using remotely sensed data to objectively identify landscape features at multiple spatial scales. The thesis focused on the potential for the use of specific digital image processing techniques to objectively generate multi-scale landscape heterogeneity models from Landsat Thematic Mapper (TM) satellite imagery. An approach using Difference-of-Gaussian (DoG) scale-space filters was used to process TM satellite data, transformed into the Normalized Difference Vegetation Index (NDVI), to define discrete landscape units across a range of scales. A series of Difference-of-Gaussian filters, with standard deviation (a) pairs of 0.9 and 1.1, 0.8 and 1.2, and 0.7 and 1.3, was applied to TM data. Results were compared against published ecoregion classifications and aerial orthophotography. These comparisons showed that the applied filters failed to accurately define landscape elements at a range of spatial scales. Filter size was found to be highly sensitive to the resolution of the image and object boundaries were maldelineated when complex boundary combinations were encountered. A modified approach using Gaussian/Difference-of-Gaussian or GaussianlLaplacian filter combinations and an alternate approach of watershed segmentation were recommended as potential techniques to test in further study. Future study should also include statistical approaches to aggregate zones and relate delineated zones to an actual, on-the-ground physiognomic classification scheme. Close