The Point Cloud, AWARE’s Electronic Newsletter
Vol. 4, Issue 3. Date: 12/21/2018
aware.forestry.ubc.ca Previous Issue
Not displaying correctly?
View this email in your browser
Happy Holidays! First of all, congratulations to Rebecca Wylie, who, just last week, successfully defended her thesis on Q14. AWARE also has just added another researcher. We would like to welcome Van-Tho Nguyen, who will be working on Q19 with Richard Fournier.
Curtis Marr, the project manager for AWARE, will be leaving to take a new position as the industrial liaison officer for Applied Science at Simon Fraser University. He will be replaced by Gladys Tecson, who has extensive experience administering research projects in UBC’s Wood Science department.
Ayla Brombach is an MESc student studying under Dr. Jeff Dech in the Forest Resources Laboratory at Nipissing University, where she completed her BSc Honours Specialization degree in biology. Ayla is currently working on Question 9 under Theme 2 of the AWARE project.
The focus of Question 9 was derived from recent studies that made use of ecological land-classification (ELC) systems for modelling wood attributes. The results of these studies suggested that broad-scale patterns of moisture availability can be linked to characteristics such as wood density and fibre length. However, these ELC systems are specific to a given region so a better approach would be to use a mechanism, in this case moisture, driving patterns of wood quality characteristics on the landscape, and to develop a classification that captures the variation in this driving force. Hence, the objective of this study is to develop predictive models of soil moisture availability from DEM data for plots across Newfoundland with the anticipation that these models will provide a useful tool for separating forest stands by different soil moisture regimes, and that they are likely to have certain average wood quality properties.
When not at the university, Ayla enjoys working with the North Bay Heritage Gardeners to attain their goal of developing a year-round, educational, and community-based program focusing on waterfront improvement. She has also taken on the role of Participant Coach for Nipissing University’s and Canadore College’s Relay for Life. Her personal time is spent out enjoying nature. Her love of nature stems from a childhood of exploring the outdoors and her keen curiosity for the flora and fauna that surrounds everyday life. She can often be found hiking in the summer and snowshoeing in the winter with her camera in tow.
Ayla plans to pursue a research career and potentially a PhD after finishing her master’s degree in the summer of 2019.
Q18 Assessing the effectiveness of ALS for identifying how forest structure affects bird abundance
Sam Herniman has completed data analysis for the first half of question 18. This question deals with modeling biodiversity over differing scales. As species extinctions are greater than the natural background rate, it is important to model the suitable habitat of biodiversity indicators on large and small scales (Pimm & Raven, 2000). The work for this question has been carried out in collaboration with Corner Brook Pulp and Paper Ltd, based in Newfoundland.
The initial goal of this question has been to figure out which aspects of the Newfoundland forests have the greatest influence on bird abundance and to determine if ALS is appropriate to measure it. This process involved Natural Resource Canada establishing enhanced forest industry (EFI) plots in the Harry’s River Watershed in western Newfoundland. Harry’s River is a 900 square kilometer watershed south west of Corner Brook. Airborne laser scanning (ALS) data was concurrently collected in the Harry’s River Watershed in 2016 so that we could compare data collected by traditional forest mensuration methods (from the EFIs) and ALS.
In July 2018 Sam Herniman traveled to Newfoundland to conduct avian point counts at the enhanced forest inventory plots in Harry’s River. This fieldwork took place with the help of Jenna McDermott, an MSc student at Memorial University of Newfoundland, and the team at Corner Brook Pulp and Paper. This fieldwork gleaned 30 point counts in which the number of each bird species present was counted.
These three data acquisitions have been used to model bird abundance in the Harry’s River Watershed using best subset regression. By modeling the abundance of each species with this method using the different types of data available (traditional plot metrics, ALS metrics, and spectral metrics) we were able to determine the goodness of fit of each method.
Figure 1 shows the mean normalized RMSE of the different data types we considered. The range bars show the maximum and minimum normalized RMSE values. Normalized RMSE is better when it is lower. The variability of ALS models and spectral models is much higher than the variability of traditional plot models, showing that plot models are more consistent than remote sensing methods but the best remotely sensed models perform better than the best plot model.
When we consider plot models as our benchmark, this shows that there is potential for combined models with ALS and spectral data to be more efficient than manual efforts. This knowledge will lead to future research into how to model bird diversity over the entirety of Newfoundland.
Figure 1. We modeled the presence-absence of each avian species found in Newfoundland using best subset regression. The mean normalized RMSE of each data type is indicated by the columns and the minimum an maximum values are illustrated by range bars.
At Laval, JR Roussel’s lidr package underwent a major re-engineering over the last ten months. With more capabilities and improved integration in the R ecosystem, the latest upgrade is expected to be ready for release on Jan. 1, 2019. Out west, UBC is making good progress on FOSTER, a set of tools for scaling up transects. We expect to release a beta version of this to be available to AWARE researchers in late January. A bit further out, UQAM’s tree delineation software (SEGMA) is also adding some new features and will likely be released in mid-late 2019.