Fall 2023 updates: Innovative technologies and data-driven adaptation

To continue our fall 2023 updates, the Silva21 community met on Tuesday, October 17th the to discuss ongoing research projects under the theme of innovative technologies and data-driven adaptation. We heard from our industry partners at Domtar and eight HQPs on multiple research projects! Check out summaries of their updates below.

---

LiDAR EFI for growth projections: new approaches - José Riofrío - Postdoctoral Fellow, UBC

Forest Resource Inventories (FRIs) require a systematic and periodic update to track forest resource changes. Remote sensing technologies, including airborne laser scanning (ALS) and satellite imagery, can be used to update FRIs periodically, enabling the development of Enhanced Forest Inventories (EFIs) in a continuous forest inventory framework. However, non-stand replacing disturbances might affect the accurate quantification of forest attribute changes by altering forest structure, species composition and growth dynamics. This project aims to combine ALS with Permanent Sample Plot (PSP) data to identify areas where non-stand replacing disturbances are likely to trigger mortality events that might cause a shift in species composition. The project focuses on boreal mixedwoods (Romeo Malette Forest, Dog River-Matawin Forest) and Great Lakes St. Lawrence (GLSL) forests in Ontario.

The PSP database structured for tracking the individual tree status was used to calculate mortality rates at plot level in terms of basal area and number of trees lost between successive measurements. This mortality rate database constitutes the main input for evaluating the ability of ALS data to link the structural attributes and size-frequency distributions of canopy gaps to mortality patterns. Currently, various modeling approaches suitable to the nature of the data (continuous proportional values that range between 0 and 1) are under evaluation, for instance, beta regression and zero-inflated models. In a second step, developed models and wall-to-wall forest attributes derived from ALS will be employed to identify areas within EFI where mortality events resulting from non-stand replacing disturbances.

José Riofrío

Postdoctoral Fellow

University of British Columbia

Supervisor: Nicholas Coops

Project Page

jose.riofriosalazar@ubc.ca

New techniques for measurements within stands - Lukas Olson, MSc student, UBC

Forest ecosystems, which cover 35% of Canada and 26% of the Earth's landmass, are crucial for global biodiversity, the economy, and local communities. In Canada, forests are significant contributors to the economy, providing employment for over 200,000 people and contributing over CAD 34 billion to our annual GDP. Globally, the forest industry is responsible for over USD 450 billion in trade and sustains over 30 million jobs. However, Canadian forests face multiple challenges, including varying disturbances and threats intensified by climate change, emphasizing the need for comprehensive monitoring and data-driven management. While there have been advancements in assessing forest health, there's still a need for efficient tools to evaluate tree-level health.

This project seeks to augment and improve the toolkit available to forest health professionals, focusing on the assessment of Sugar Maple health with potential applications to various other deciduous broadleaf species. It employs drone-based intracanopy photogrammetry to record videos of entire trees. This footage undergoes photogrammetric processing, a technique that transforms a series of photos into 3D point clouds. The aim of this project is to deduce quantitative health metrics such as branch density, bark conditions, crown density, crown status, and epicormic sprouting presence to calculate tree vigour.

While the project has addressed the evaluation of structural components like stems and branches, a prominent challenge arises from the potential movement of canopy leaves due to wind and droneinduced airflows. These factors can introduce noise and distortions during photogrammetric reconstruction, making further processing of the generated point cloud difficult and less accurate under traditional methods.

To navigate these challenges, the project is adopting a composite approach. It uses quantitative structural models for the tree's structural parts, which are less likely to move, and employs voxel-based Canopy Analysis to calculate crown density, status, and overall foliage health. This method maps point cloud data onto a 3D grid, attributing specific traits to individual grid cells. Subsequently, these cells are categorized as either Structural or Foliage, with foliage point density and colour used for further analysis, offering a richer perspective on tree health.

The potential influence of this research on sustainable forest management is significant. By devising a data-driven and economical method for assessing tree health, the myriad benefits forests offer - economic, ecological, and cultural - could be better managed. More broadly, this research could contribute to the maintenance of the long-term health and productivity of Canada's forests, thereby promoting sustainable forest management and conservation initiatives globally.

Lukas Olson

MSc student

University of British Columbia

Supervisor: Nicholas Coops

Project page

olson@student.ubc.ca

Using Planetscope to characterize disturbances in Canada's boreal forest - Spencer Shields, MSc student, UBC

Spatially-explicit forest inventories which are up to date in near-real time would be invaluable tools for forest management. One piece of the puzzle in developing these is being able to quickly detect forest disturbances and evaluate their impact. The focus of my research is finding a way to do this using a network of Earth observation satellites called Planetscope.

Planetscope is a constellation of about 150 micro-satellites that orbit the Earth and produce 3m optical imagery of its entire surface every single day. This is a huge improvement in spatial and temporal resolution over satellites such as Landsat or MODIS which have traditionally been used for forest monitoring. However, Planetscope data can be challenging to work with because there are often inconsistencies between the sensors in its many, many satellites. This makes analyzing their data challenging because you might not be able to tell if a change you see between two images is caused by a change in the forest or the fact that they were taken by different satellites.

In this update I present one potential solution to this issue, which is z-score normalization. This is accomplished by calculating a vegetation index (such as NDVI) for each image; masking out parts of the image that do not represent forest; then calculating the z-score for each pixel taking the pixel’s value, subtracting the average pixel value for the entire image, and dividing this by the standard deviation of each pixel value in the image. The logic behind this method is that you go from looking at an absolute value for each pixel to the relative value of that pixel versus the other pixels in its image: the former might change due to differences in sensors, but the latter will only change due to changes in what is being imaged. For this reason, other researchers have used this method for combining and comparing satellites images from lots of different sources.

I have found that this method works well for correcting the differences between Planetscope sensors and detecting forest disturbances. A next step will be to evaluate how well this method works for evaluating the severity of forest disturbances.

Spencer Shields

MSc student

University of British Columbia

Supervisor: Nicholas Coops

Project page

spenshi@student.ubc.ca

Viability of climate-informed landscape-level management strategies - Kirk Johnson, PhD student, UBC

The forests near Quesnel, British Columbia have experienced major disturbances over the last twenty years. The mountain pine beetle (Dendroctonus ponderosae) epidemic caused extensive pine mortality in the mid-2000’s and early 2010’s, significantly reducing the amount of mature timber. In addition, large wildfires burned a substantial portion of the area’s harvestable timberland in 2017. Due to these losses, sustainable harvest levels will be curtailed over the next 50 years, assuming management under a traditional clearcut/plant silvicultural regime. Climate change may further compound these management challenges and negatively impact critical timber species like lodgepole pine and hybrid spruce.

To address these issues, Kirk’s project under Silva21 (AN.8b) will explore climate-informed management at the landscape scale, incorporate climate risk into forest estate models, and identify landscape-level silvicultural strategies that foster resilience and production. Silvicultural treatments like precommercial and commercial thinning have the potential to accelerate rotations, increase piece size, and help ameliorate climate stress. Thinning may also soften timber supply shortages and help supplement traditional management techniques. Given these possible benefits, Kirk’s project will investigate thinning as an adaptive silvicultural strategy in the Quesnel area. Kirk will also include climate considerations like long-term species suitability and drought potential when selecting sites for thinning in landscape-level forest estate models. When complete, this project could help inform silvicultural decisions and sustainable harvest levels in the Quesnel area, with implications for timber management across western Canada.

Kirk Johnson

PhD student

University of British Columbia

Supervisor: Nicholas Coops

Project page

kmj1@student.ubc.ca

Wood properties as proxies for past climate conditions - Philippe Riel, MSc student, Université Laval

My project aims to find an alternative indicator to characterize climate events such as droughts, different from tree ring width. Certain wood properties are sensitive to the growth conditions in which trees develop, such as microfibril angle, which is particularly sensitive to variations in water availability. Therefore, the objective is to verify whether the microfibril angle in black spruce can serve as a reliable indicator for characterizing droughts. Microfibril angle is measured at the fine scale of an annual growth ring to associate these measurements with annual water availability. The technique used to obtain microfibril angle is the polarized Raman spectroscopy technique, which determines the orientation of crystalline cellulose in the wood cell walls.

It was difficult to obtain such equipment on the campus of Université Laval, but we came up with a solution through a partnership with a professor in the chemistry department of the Université de Montréal. This allows us to measure the microfibril angle of cellulose on my black spruce samples, while having access to the advice and assistance of spectroscopy specialists who use the equipment frequently. This laboratory measurement period is expected to last a total of 4 to 5 weeks.

We are looking to take microfibril angle measurements in earlywood and latewood through the same few years for each tree. The samples came from black spruce trees harvested at two different sites in the Canadian boreal forest. One site near Timmins in Ontario and another site in the Lac-Saint-Jean region in the province on Québec.

Philippe Riel

MSc student

Université Laval

Supervisor: Alexis Achim

Project Page

philippe.riel.1@ulaval.ca

---

If you are a member of the Silva21 team and would like to receive a copy of all slides, log in to our Members area.

Forgot the password? Email amy.wotherspoon@ubc.ca.

Our next update meetings will take place in Spring 2024! Stay tuned with all news Silva21 by subscribing to our newsletter at the bottom of this page.

---