Source: UNIVERSITY OF MAINE submitted to NRP
REMOTE SENSING OF FOREST ENVIRONMENTS
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
MCINTIRE-STENNIS
Reporting Frequency
Annual
Accession No.
0229729
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Oct 1, 2012
Project End Date
Dec 31, 2015
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
UNIVERSITY OF MAINE
(N/A)
ORONO,ME 04469
Performing Department
School of Forest Resources
Non Technical Summary
As Maine forest landscape continues to change in response to economic and societal expectations, it has become critical for researchers and natural resource managers to have access to the most advanced information and technologies to plan for the future. Research integrating remote sensing, spatial statistics and modeling will explore the Maine forest vulnerability to spruce budworm defoliation, and watershed level forest fragmentation trends from two decades, past to the present. Methods to estimate forest biomass over large forest landscapes will be needed for preparing Maines forest sector for sustainability in future carbon markets. The objectives are important to the state of Maine in working toward improved methods of forest resource assessment using appropriate remote sensing technology.The project results will be of interest to municipal planners and major landowners to understand how forests are changing and how changes can be monitored and resource impacts can be predicted into the future.
Animal Health Component
(N/A)
Research Effort Categories
Basic
100%
Applied
(N/A)
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1230610107010%
1230613107010%
1230620107010%
1237210107060%
1367210107010%
Knowledge Area
136 - Conservation of Biological Diversity; 123 - Management and Sustainability of Forest Resources;

Subject Of Investigation
0610 - Conifer forests of the North; 0620 - Broadleaf forests of the North; 7210 - Remote sensing equipment and technology; 0613 - Mixed conifer-broadleaf forests;

Field Of Science
1070 - Ecology;
Goals / Objectives
Goal: The proposed research will use traditional remote sensing technologies (e.g., Landsat Thematic Mapper (TM) and high spatial resolution imagery) and emerging technologies (Light Detection and Ranging (LIDAR) and Synthetic Aperture Radar (SAR)) to develop improved algorithms and data fusion techniques to produce maps of various forest attributes; like canopy cover type, species composition, stand age or size class, disturbance rate and intensity, and above-ground biomass,to be generated on a regular basis. Objectives: Perform remote sensing investigations of: 1.Forest Change Dynamics and Prediction of Spruce Budworm Vulnerability,(a) Quantify recent forest harvesting rates and harvest intensity on commercial forestlands in Maine using a Landsat-derived time series of forest disturbance, (b) Calibrate the forest landscape model LANDIS-II to northeastern tree species assemblages, especially those susceptible to spruce budworm defoliation, (c) Develop maps of current forest conditions, including composition, age structure, and spruce budworm vulnerability across 4 million hectares of commercial forestland in Maine.2.Forest Structure, Above-ground Biomass, and Carbon Estimation,(a) Develop disturbance ecology scenarios and modeling applications by combining Landsat time-series imagery and LANDIS-II landscape simulation model to predict future forest composition and biomass, (b) evaluate the potential of emerging remote sensing technologies to predict forest structure attributes, biomass, and carbon stocks (c) develop predictive relationship between remote sensing metrics and ground-based sampling plots; and (d) produce large regional maps that display key forest attributes, particularly above-ground biomass. 3.Changing Forest Landscape Composition and Configuration,(a)Select appropriate landscape patch metrics based on literature review and experimentation; (b) Quantify the extent of forest fragmentation of mature forest including pattern, and trends from the early 1990s to 2007 by Maine watersheds and, contrast differences among landowner type, between northern Maine unorganized townships and southern Maine organized townships, and Maine ecoregions; (c) Develop predictor variables and maps of large diameter forests in northern Maine, through integration of historical change maps, GIS data, and interpretation of recent high resolution imagery. The project results will be of interest to municipal planners and major landowners to understand how forests are changing and how changes in forest type, structure, and biomass can be monitored to predict resource impacts into the future. This can lead to improved planning to serve the community and public.
Project Methods
Forest Change Dynamics and Prediction of Spruce Budworm Vulnerability: Time-series medium spatial resolution satellite imagery will be combined with FIA data and spatial landscape disturbance models to predict and map the vulnerability of northern forest stands to spruce budworm defoliation. The forest disturbance maps are being developed using time-series change detection methods developed over the past two decades. Spatial vulnerability models will be developed based on known relationships with host (e.g., balsam fir and spruce spp., % live tree basal area) and non-host species relative abundance and forest age, using a partial least squares regression procedure and a genetic algorithm for optimal selection of satellite-derived predictor variables. Maps of fir and host species abundance will be combined with satellite-derived forest age maps to produce a preliminary map of budworm vulnerability, depicting the spatial distribution of vulnerability classes defined by species composition and maturity. Results will be used to evaluate budworm outbreak impact using an existing spruce budworm decision support system. Forest Structure, Above-ground Biomass and Carbon Estimation: Landsat time-series disturbance maps will be combined with the LANDIS landscape projection system to forecast changes in forest composition, structure, and biomass/carbon dynamics.FIA plot data will be examined to develop height and canopy density correlations with LIDAR and synthetic aperture radar data. We expect to make methodological advances in combining LIDAR, Radar and spatio-temporal forest composition and structure data developed from Landsat time-series imagery to quantify forest biomass and carbon stocks. Modeling applications and disturbance ecology scenarios (LANDIS-II) will be used to predict future forest composition and biomass. Forest Landscape Composition, Configuration,and Metrics: Using fragmentation metrics and statistical analyses (e.g., principal components analysis), this study will quantify differences in landscape fragmentation pattern and trends among Maine ecoregions and their watersheds and, between the industrial, corporate-owned northern forest and the more developed family-owned and municipal southern forest landscape of Maine. Metrics that are standardized by area will used to account for varying watershed size. Output from the fragmentation analysis will be analyzed by using Principal Components Analysis to determine how the landscape configuration changed due to forest harvesting over the study period. Variance in landscape change due to geographic location will also be examined by using a lake categorization map of Maine, which was developed from a number of biophysical variables to determine three major ecoregions in the state.

Progress 10/01/15 to 09/30/16

Outputs
Target Audience: Nothing Reported Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Nothing to report due to retirement December 31, 2015.

Publications


    Progress 10/01/12 to 12/31/15

    Outputs
    Target Audience:Research cooperators and stakeholders include various forest management companies in Maine (through the Maine Cooperative Forestry Research Unit), Maine Department of Inland Fisheries and Wildlife, NASA Stennis Space Center, U.S. Forest Service-Northern Research Station - FIA Unit, U.S. Geologic Survey - Maine Cooperative Fish and Wildlife Unit, and New England Forestry Foundation. The assortment of research cooperators and landowner types is broadly representative of the unorganized townships of northern Maine where most of our intensive studies take place. Changes/Problems:Due to my retirement 2 years prior to the planned termination date of the project, there were some objectives of the major goals not accomplished in the shorter time period. For example, objective 1.2 d. Forest Structure, we did not produce regional maps for forest composition and biomass and objective 3 b & c., Changing Forest Landscape Composition, we did not successful publish forest frafmentation studies and comparisons across Maine regions and watersheds (although some research was attempted and terminated due to some technical problems and poor results).. However, there were minor changes in objectives due to changes in focus (influenced by funding) that occurred after the project was approved and initiated..We intiated research of water clarity of Maine lakes detected remotely by two types of orbiting satellites (Landsat and MODIS). We developed regression relationships through correlating in situ water column clarity (using secchi disk) to satellite multi-spectral reflectance and combined geographic data integration and to develop equations to predict water clarity for Maine lakes (see several publication products resulting from this research). What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?Through journal publications and conference presentations. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

    Impacts
    What was accomplished under these goals? We have developed methods to map spruce budworm host species distributions and forest vulnerability classes using Landsat satellite imagery and forest attribute data provided by USFS Forest Inventory and Analysis (FIA) field plots. Budworm vulnerability classes were defined based on known relationships with host and non-host species relative abundance and forest age. Host species relative abundance estimates (expressed as a proportion of total live aboveground biomass) were produced using a predictive modeling framework based on advanced machine learning algorithms. Our approach overcomes multiple deficiencies of more established methods, most notably the tendency to over-predict abundance at low values and under-predict abundance at high values. We have utilized the same approach to predict and map forest disturbance from the nearly 40-year archive of Landsat satellite imagery. Forest disturbance data were combined with host species distributions tomap budworm vulnerability. Additionally, satellite-derived maps were integrated with aspatial FIA data to develop a stand map schema that can be applied uniformly across multiple ownerships to model the outcomes of alternative management strategies and budworm outbreak scenarios. Satellite-derived proxy stand maps were integrated with forest estate planning software and the Spruce Budworm Decision Support System (MacLean et al. 2001; Hennigar et al. 2011) to evaluate the outcomes of alternative outbreak and response scenarios across a 300,000 acre portion of our study area.

    Publications

    • Type: Conference Papers and Presentations Status: Published Year Published: 2012 Citation: Legaard, K.R., S. Sader, J. Wilson, E. Simons-Legaard, and A. Weiskittel. 2012. Mapping vulnerability to defoliation by spruce budworm using Landsat satellite imagery and FIA field plots. Poster presented at the Eastern CANUSA Forest Science Conference, Durham, New Hampshire
    • Type: Conference Papers and Presentations Status: Published Year Published: 2012 Citation: Simons-Legaard, E., K. Legaard, J. Wilson, A. Weiskittel, and S. Sader. 2012. Long-term outcomes and tradeoffs of forest policy and management practices on the borad-scale sustainability of forest resources: wood supply, carbon, and wildlife habitat. Poster presented at the Eastern CANUSA Forest Science Conference, Durham, New Hampshire.
    • Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: Legaard, K.R., S. Sader, J. Wilson, E. Simons-Legaard, and A. Weiskittel. 2013. A spatial assessment of vulnerability to defoliation by spruce budworm across the commercial forestland of northern Maine. Presented at the New England Society of American Foresters Spring Meeting, Bethel, Maine.
    • Type: Journal Articles Status: Published Year Published: 2014 Citation: Hayashi, R., A. Weiskittel and S.Sader 2014. Assessing the feasibility of low density LiDAR for stand inventory attribute predictors in complex and managed forests of northern Maine, USA. Forests:(5) 363-383.
    • Type: Journal Articles Status: Published Year Published: 2015 Citation: Legaard, K., S.A. Sader, and E. Simons. 2015. Evaluating the impact of abrupt changes in forest policy and management practices on landscape dynamics: analysis of a Landsat image time series in the Atlantic Northern Forest. PLOS. PLOS ONE | DOI:10.1371/journal.pone.0130428 June 24, 2015
    • Type: Journal Articles Status: Published Year Published: 2015 Citation: Dunckel, K., A. Weiskettel, G. Fiske, S. Sader, E. Latty and A. Arnett. 2015. Linking remote sensing and various site factors for predicting the spatial distribution of eastern hemlock occurrence and relative basal area in Maine, USA Forest Ecology and Management:(358) 180-191.
    • Type: Journal Articles Status: Published Year Published: 2012 Citation: Noone M.D. and S.A. Sader. 2012. Are forest disturbances influenced by ownership change conservation easement status, and land certification? Forest Science 58:119-129.
    • Type: Journal Articles Status: Published Year Published: 2013 Citation: McCullough, I.M., C.S. Loftin, and S.A. Sader. 2013. Lakes without Landsat? An alternative approach to remote lake monitoring with MODIS 250 imagery. Lake and Reservoir Management 29(2): 89-98.
    • Type: Journal Articles Status: Published Year Published: 2012 Citation: Sader, S.A. and T. Mueller 2012. 2011 survey on remote sensing and geospatial technologies training in U.S. and Canadian colleges and universities. Photogrammetric Engineering and Remote Sensing 78(12): 1193-1198.
    • Type: Journal Articles Status: Published Year Published: 2012 Citation: McCullough, I.M., C.S. Loftin, and S.A. Sader. 2012. Combining lake and watershed characteristics with Landsat TM data for remote estimation of regional lake clarity. Remote Sensing of Environment. Remote Sensing of Environment 123 (2012) 109115
    • Type: Journal Articles Status: Published Year Published: 2012 Citation: McCullough, I.M., C.S. Loftin, and S.A. Sader. 2012. High-frequency remote monitoring of large lakes with MODIS 500 m imagery. Remote Sensing of Environment 124: 234-241


    Progress 10/01/14 to 09/30/15

    Outputs
    Target Audience:Research cooperators and stakeholders include various forest management companies in Maine (through the Maine Cooperative Forestry Research Unit), Maine Department of Inland Fisheries and Wildlife, NASA Stennis Space Center, U.S. Forest Service-Northern Research Station - FIA Unit, U.S. Geologic Survey - Maine Cooperative Fish and Wildlife Unit, and New England Forestry Foundation. The assortment of research cooperators and landowner types is broadly representative of the unorganized townships of northern Maine where most of our intensive studies take place. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?As mentioned above, the SBW stand vulnerability maps have been distributed to the Maine Forest Service and landowners for pre-emptive management intervention and monitoring of SBW moth populations in Maine. Project results have been reported in manuscripts published in peer-reviewed journals, MS theses and conference presentations. What do you plan to do during the next reporting period to accomplish the goals?A proposal was submitted to the Northern States Research Cooperative program to continue research on forest change mapping and SBW outbreak but was not selected for funding, presumably because the SBW outbreak had not yet occurred in Maine. The proposal will be revised and resubmitted when SBW infestation is first reported in Maine. The PI is retiring on January 1, 2016. A final report on this project will be submitted during the first quarter of 2016.

    Impacts
    What was accomplished under these goals? Most of the accomplishments this period have focused on Objective 1 (forest change dynamics) and objective 3 (Changing forest landscapes). A SBW stand vulnerability map (developed by this project and reported in an earlier report) were made available to the Maine Forest Service and major forest landowners in northern Maine. The maps were used to determine locations of high vulnerability stands where pheromone traps were positioned all over northern and eastern Maine unorganized townships. SBW defoliation is active in southern Quebec and some activity has been detected in New Brunswick, Canada. In Maine, high vulnerability stands are being monitored closely and although adult SBW moth population counts have been increasing over the past 3 years, no significant larval feeding has yet been detected inside the Maine border. Under objective 2, LiDAR research conducted in the Penobscot Experimental Forest has been completed. This research resulted in a completed MS degree in 2014 and an article published in a peer-reviewed journal. Research concerning Objective 1 & 3 resulted in a peer reviewed article published in PLOS ONE about the impacts of current forest harvesting practices that are difficult to disentangle from the persistent influences of past practices. Using a time series of Landsat satellite imagery (1973-2010) we assessed cumulative landscape change caused by these two very different management regimes. We modeled predominant temporal patterns of harvesting and segmented a large study area into groups of landscape units with similar harvest histories. Time series of landscape composition and configuration metrics averaged within groups revealed differences in landscape dynamics caused by differences in management history. In some groups (24% of landscape units), salvage caused rapid loss and subdivision of intact mature forest. Persistent landscape change was created by large salvage clearcuts (often averaging > 100 ha) and conversion of spruce-fir to deciduous and mixed forest. In other groups little affected by salvage (56% of landscape units), contemporary partial harvesting caused loss and subdivision of intact mature forest at even greater rates. Management regimes impacted different areas to different degrees, producing different trajectories of landscape change that should be recognized when studying the impact of policy and management practices on forest ecology.

    Publications

    • Type: Journal Articles Status: Published Year Published: 2015 Citation: Legaard, K., S.A. Sader, and E. Simons. 2015. Evaluating the impact of abrupt changes in forest policy and management practices on landscape dynamics: analysis of a Landsat image time series in the Atlantic Northern Forest. PLOS. PLOS ONE | DOI:10.1371/journal.pone.0130428 June 24, 2015


    Progress 10/01/13 to 09/30/14

    Outputs
    Target Audience: Research cooperators and stakeholders include various forest management companies in Maine (through the Maine Cooperative Forestry Research Unit), Maine Department of Inland Fisheries and Wildlife, NASA Stennis Space Center, U.S. Forest Service-Northern Research Station - FIA Unit, U.S. Geologic Survey - Maine Cooperative Fish and Wildlife Unit, and New England Forestry Foundation. The assortment of research cooperators and landowner types is broadly representative of the unorganized townships of northern Maine where most of our intensive studies take place. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? The SBW stand vulnerability maps have been distributed to the Maine Forest Service and landowners for pre-emptive monitoring of SBW moth populations in Maine.Project results have been reported in manuscripts submitted and published in peer-reviewed journals, MS theses and conference presentations. What do you plan to do during the next reporting period to accomplish the goals? Two pre-proposals were submitted to the Northern States Research Cooperative program to continue research on forest change mapping and SBW outbreak, spread and stand vulnerability. Under current funding through NSRC and the NASA/USDA National Institute of Food and Agriculture (NIFA) Carbon Cycle Science program, we are currently applying the LANDIS-II (LANDscape Disturbance and Succession) simulation model over our entire 10 million acre northern Maine study area. Alternative forest management, budworm outbreak, and climate change scenarios will be executed to investigate impacts on wood supply, carbon stocks, and wildlife habitat.

    Impacts
    What was accomplished under these goals? A SBW stand vulnerability map was developed under two U. S. Forest Service, Northern State Research Cooperative funded research projects. The maps were made available to the Maine Forest Service and major forest landowners in northern Maine. The maps are being used to determine locations of high vulnerability stands where pheromone traps will be positioned all over northern and eastern Maine unorganized townships. Forest change detection maps support the SBW vulnerability mapping research. One manuscript on change detection methods was submitted to a peer-reviewed journal in October 2014. LiDAR research was conducted in the Penobscot Experimental Forest and an article was published in a peer-reviewed journal. A graduate student leading this research completed his master of science degree in 2014.

    Publications

    • Type: Journal Articles Status: Published Year Published: 2014 Citation: Hayashi, R., A. Weiskittel and S.Sader 2014. Assessing the feasibility of low density LiDAR for stand inventory attribute predictors in complex and managed forests of northern Maine, USA. Forests:(5) 363-383.


    Progress 10/01/12 to 09/30/13

    Outputs
    Target Audience: Research cooperators and stakeholders include various forest management companies in Maine (through the Maine Cooperative Forestry Research Unit), Maine Department of Inland Fisheries and Wildlife, NASA Stennis Space Center, U.S. Forest Service-Northern Research Station - FIA Unit, U.S. Geologic Survey - Maine Cooperative Fish and Wildlife Unit, and New England Forestry Foundation. The assortment of research cooperators and landowner types is broadly representative of the unorganized townships of northern Maine where most of our intensive studies take place. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? Through journal publications and conference presentations What do you plan to do during the next reporting period to accomplish the goals? Under funding obtained through NSRC and the NASA/USDA National Institute of Food and Agriculture (NIFA) Carbon Cycle Science program, we are currently applying the LANDIS-II (LANDscape Disturbance and Succession) simulation model over our entire 10 million acre northern Maine study area. LANDIS-II simulates disturbance and succession of cohorts principally defined by age and species. In order to produce suitably detailed spatial data, we have modeled and mapped the distributions of 13 economically and ecologically important tree species and integrated these results with forest disturbance products. Alternative forest management, budworm outbreak, and climate change scenarios will be executed to investigate impacts on wood supply, carbon stocks, and wildlife habitat.

    Impacts
    What was accomplished under these goals? We have developed methods to map spruce budworm host species distributions and forest vulnerability classes using Landsat satellite imagery and forest attribute data provided by USFS Forest Inventory and Analysis (FIA) field plots. Budworm vulnerability classes were defined based on known relationships with host and non-host species relative abundance and forest age. Host species relative abundance estimates (expressed as a proportion of total live aboveground biomass) were produced using a predictive modeling framework based on advanced machine learning algorithms. Our approach overcomes multiple deficiencies of more established methods, most notably the tendency to over-predict abundance at low values and under-predict abundance at high values. We have utilized the same approach to predict and map forest disturbance from the nearly 40-year archive of Landsat satellite imagery. Forest disturbance data were combined with host species distributions to map budworm vulnerability. Additionally, satellite-derived maps were integrated with aspatial FIA data to develop a stand map schema that can be applied uniformly across multiple ownerships to model the outcomes of alternative management strategies and budworm outbreak scenarios. Satellite-derived proxy stand maps were integrated with forest estate planning software and the Spruce Budworm Decision Support System (MacLean et al. 2001; Hennigar et al. 2011) to evaluate the outcomes of alternative outbreak and response scenarios across a 300,000 acre portion of our study area.

    Publications

    • Type: Conference Papers and Presentations Status: Other Year Published: 2012 Citation: Simons-Legaard, E., K. Legaard, J. Wilson, A. Weiskittel, and S. Sader. 2012. Long-term outcomes and tradeoffs of forest policy and management practices on the borad-scale sustainability of forest resources: wood supply, carbon, and wildlife habitat. Poster presented at the Eastern CANUSA Forest Science Conference, Durham, New Hampshire.
    • Type: Journal Articles Status: Published Year Published: 2013 Citation: McCullough, I.M., C.S. Loftin, and S.A. Sader. 2013. Lakes without Landsat? An alternative approach to remote lake monitoring with MODIS 250 imagery. Lake and Reservoir Management 29(2): 89-98.
    • Type: Journal Articles Status: Published Year Published: 2013 Citation: McCullough, I.M., C.S. Loftin, and S.A. Sader. 2013. Application of Landsat TM imagery reveals declining water clarity of Maines lakes during 1995 to 2010.. Freshwater Science 32(3):741-752
    • Type: Other Status: Published Year Published: 2013 Citation: McCullough, I.M., C.S. Loftin, and S.A. Sader. 2013. A manual for remote sensing of Maine lake clarity. Maine Agricultural and Forestry Experiment Station, Technical Bulletin 207. 23 pp.
    • Type: Conference Papers and Presentations Status: Published Year Published: 2012 Citation: Legaard, K.R., S. Sader, J. Wilson, E. Simons-Legaard, and A. Weiskittel. 2012. Mapping vulnerability to defoliation by spruce budworm using Landsat satellite imagery and FIA field plots. Poster presented at the Eastern CANUSA Forest Science Conference, Durham, New Hampshire.
    • Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: Legaard, K.R., S. Sader, J. Wilson, E. Simons-Legaard, and A. Weiskittel. 2013. A spatial assessment of vulnerability to defoliation by spruce budworm across the commercial forestland of northern Maine. Presented at the New England Society of American Foresters Spring Meeting, Bethel, Maine.