Forest Investment Account (FIA) - Forest Science Program
FIA Project M086016

    Mountain Pine Beetle Red Attack Shelf Life Discriminations
 
Project lead: Roberts, Arthur (Simon Fraser University)
Contributing Authors: Roberts, Arthur; Bone, Christopher; Dragicevic, Suzana; Ettya, Aviv; Northrup, James; Reich, Richard W.; Becker, T.
Subject: Forest Investment Account (FIA), British Columbia
Series: Forest Investment Account (FIA) - Forest Science Program
Description:
Since cutting “bug wood” has the highest current priority any detection and mapping procedure that will help discriminate new red attacked trees from older “reds” has important implications for determining shelf life of the dead timber. These mapped mortality discriminations will assist with harvest planning and scheduling and will have implications for decisions involving timber yield and cost effective recovery of dead timber. Satellite imagery is very good for general overview evaluations but suffers from: (1) availability problems due to cloud cover & orbital periodicity, and (2) inadequate spatial resolution (often sub pixel for individual crowns) for forest health issues that require a statistically suitable spectral image for individual tree crown level analysis (Franklin et. al., 2003; Leckie et. al., 2001; Skakun et. al., 2003; Wulder and Dymond, 2004). The use of airborne imaging spectrometers (e.g. CASI) and multispectral scanners (e.g. GERS) has proved disappointing (Heath 2001; Roberts et.al., 2003). There have been no satisfactory orbital results for reliable early detection or operational red attack monitoring of MPB and other forest pests. Secondly, airborne "line-imaging" systems are not good mapping instruments in comparison with aerial photography. This inevitably drives up the costs for any operational project. Other than one study in the early 1980’s (Gimbarzevsky et. al., 1992), an obvious gap in remote sensing research, related to these forest health problems, has been the almost complete lack of experimentation (and comparison) with digitally converted colour, colour infrared and multispectral aerial photography. In particular it is recommended that all experimental studies be evaluated against this standard in terms of: (i) mapping accuracy, (ii) potential for automated spectral interpretations, and, (iii) cost effectiveness. Aerial photography is the “gold standard” in terms of mapping. Six areas of mountain pine beetle (MPB) infestation (to the west and southwest of Prince George) have been sequentially imaged, by the SFU Remote Sensing Laboratory, from April 2002 to October 2003, 2005 and from April to October 2006. These sites were established by MoF as test sites for the evaluation of experimental remote sensing systems and procedures. The extent of MPB current attack, red attack and healthy trees was ground truthed (mapped) at the individual tree level by field examination in the fall and winter of 2001 and subsequently 2002. All new red attacked trees were identified on 2002 and subsequently on 2003 digital multispectral aerial photography of these sites. There is a clearly detectable difference between the new red attacked trees mapped in Aug/Sept 2002 and the previous red attacked trees identified by ground truth in Nov/Dec 2001 (see Figure 1 attached). Comparable distinctions are also apparent on our 2006 imagery. These difference are quite distinct using four-band (B, G, R, NIR) multispectral imagery and 2006 results show that these distinctions are also in applications using combined colour and colour IR aerial photography mapping packages. Such capability is available in the Canadian private sector aerial survey industry and could be implemented for competitive tendering. Our 2003, 2005 and 2006 imagery for these same study sites confirmed this discrimination of new vs. old reds. Clearly the next step is to determine strength of replication and potential to separate current (new) and earlier reds (as well as older grays). This required a systematic replication study with new imagery in 2006 and 2007. A significant change in 2006 was the imaging evidence of an extensive new MPB attack in younger lodgepole pine stands. Our proposed 2007 imagery will permit us to evaluate 2 years of mortality differences in this younger pine plantation stands as well as determining the potential mortality descrimination (shelf life trajectory) across six years of sequential imaging: will it be possible to sepatate individual years of mortality on the 2007 imagery or will some of the earlier years of kill merge into a general "older than..." category. This study involves the continuing acquisition and digital conversion of high-resolution multispectral (visible & near infrared), colour and infrared aerial photography for our six established study areas, our new young pine sites, and a longer term study area that is cooler and wetter than the six selected test sites and site areas that were imaged in 2002/3. The selected site areas are part of the Dzitlainli Nation’s traditional territory. They contain sites in a protected area (Fleming) and Canfor TFL areas managed and logged by Canfor and Jan Cho Forestry Management. These areas are currently infested with MPB and this infestation is expected to continue to spread over the next two years. The site areas were selected in consultation with Canfor and Jan Cho with a consideration for longer term monitoring potential (unlikely to be completely over-run in the next year). The proposed research will involve flying: multispectral aerial photography in four spectral image bands: blue, green, red and near infrared. The specific spectral wavelengths will be adjusted with optical filters. A “parallel” imagery data set that can be evaluated as a private sector, competitively-contracted, form of digital multispectral remote sensing image composites will include a twin camera set-up using synchronized mapping cameras to acquire normal colour and colour IR aerial and/or B&W IR photography. These synchronized images will be scanned and registered into 4 & 6 band multispectral images for enhancement, interpretation and classification. The flight frequency will be approximately every two to four weeks over the site areas. Imaging flights will commence in April 2007 and continue into late September, early October. This schedule will be a repeat or the previous years' schedules for the same areas. Replication in this subsequent year (2007) is essential to this project as it will: (1) permit further refinement and verification of the discrimination procedures, and; (2) provide a controlled 2 year across-time imaging cycle for monitoring and mapping new and old red attack under varying site conditions. A total of 20-40 (4-6 flights for each study site) imaging flights are anticipated. Data will be analyzed throughout and following this second field season and the optimum film filter, spectral band and analytical procedures for early reliable discrimination of new vs. old red attack that were determined in previous flights will be replicated in 2007. In this "shelf life trajectory" project we are undertaking same year and multi-year discriminations of new and old red attacked trees. Our initial results show that this worked quite well on our 2002/3, 2005 and 2006 data for our six previous study sites and our selected new sites. We will be replicating these results and determining the strength of this relationship for our other young pine sites. In addition it is essential to determine if this relationship will continue across multiple years: i.e. can three or more years of red attack be reliably separated and what is the shelf life performance up to gray attack? For this reason an additional field season is planned for 2007. Depending upon results, there may be some further follow up research.

    Deliverables:

Ecological Modeling Article, 192(1-2):107-125
MPB Red Attack Shelf Life Discriminations... (Report) (4.3Mb)
Evaluation of Airborne Imagery (Report) (1.0Mb)
Israeli Geographical Association Annual Conference (Hebrew Abstract) (25Kb)
Israeli Geographical Association Annual Conference (English Abstract) (91Kb)
Environmental Modeling and Assessment Article, 12(2):105-118.

To view PDF documents you need Adobe Acrobat Reader, available free from the Adobe Web Site.

Updated August 16, 2010 

Search for other  FIA reports or other Ministry of Forests and Range publications.

Please direct questions or comments regarding publications to For.Prodres@gov.bc.ca