Discussions
Christmas Bird Count Limitations
Unfortunately, the data I had retrieved from the Christmas Bird Counts surveys lacked significant spatial information. Although they contained useful about bird sightings, they were aggregated over the entire city. This limitation prevented me from performing multivariate analyses or running statistical tests through programs like FragStats. As a result of these limitations I was relegated to only analyzing trends in raw numbers.
Inconsistencies in Data Time Frames
This study was further limited by the confines of our data sets. It is important to consider that a mismatch in time frames or outdated information can lead to uncertainties and errors in our results. The Landsat satellite imagery of the area for example only had files as far back as 1999. While Christmas Bird counts dated back to the early 1900s l, I was unable to match land use changes across that same time period. If we had earlier images of the city I believe that the change in landscape would be even greater. Likewise, information about building heights was only collected in 2009 and would not reflect recent construction projects and developments. While I would have preferred to use more relevant information about the buildings in the city, this layer was crucial in assessing suitable habitats.
Visualizing Landscape Change
For the first stage of this project I chose to simple use a visual comparison of the NDVI results and not a more analytical process like Image Math because of errors that were present in the satellite images. Following 2003 when the Landsat 7 satellite began to malfunction, all images taken after this point in time had diagonal black lines that led to missing values. To solve this issue, I tried a number of techniques. An existing mask layer came packaged with the satellite image but I wasn’t able to merge the TIF files together in a way that allowed the software to accurately perform an NDVI analysis. When this became clear, I then tried to use Landsat 8 imagery in its place. Unfortunately, the image that I downloaded appeared fuzzy (likely a result from atmospheric conditions at the time) and so when I performed the NDVI assessment, the results were inconsistent to the 1999 image and could not serve as a proper comparison. Without any other option available to me, I simply overlaid the recent 2017 layer over the complete 1999 layer, considering that the difference would be minimal in comparison to a clearer image. Unfortunately, this meant that the missing values still existed in the 2017 files and so the Image math process would not yield accurate results. I decided that a visual comparison was an accurate method that would meet the needs of my project.
Bird Guilds
Separating the 250+ species of birds into their four main categories was a subjective process and was based on my limited knowledge of these organisms.
Why did the Common Birds and Predator Bird categories show an increase in bird populations when the majority of birds were declining?
While bird populations overall have declined over time in Vancouver, two categories of birds have shown signs of being capable of thriving in this unfamiliar habitat. It is important to assess why these birds have been able to survive where others have not. The Common Birds, which I like to describe as rats of the sky - have adapted their behaviours around human activities. Pigeons and crows for example scavenge the city’s trash and litter for food. Unlike their In-Land Bird counterparts who have less plastic diets, they have been able to dominate an unconventional food source. To reflect this, weighted sum analysis for this category of birds would favour high population density neighbourhoods where there is likely a larger food source. Predator Birds have also experienced a growth in numbers because the towering skyscrapers have suited their natural niches (Map ) but these trends could also be correlated to the spike in Common Birds who are easily hunted. With higher prey abundances, the increase in hunters could simply be an effect of Predator-prey relationships. While these are just a few possible solutions, understanding the processes and conditions that have allowed these particular birds to flourish is crucial in building future studies and management strategies.
Troubleshooting with the Weighted Sum Function
When performing the weighted suitability map there was issues with the individual layers that complicated the process. For example, the rapid transit lines were only around the Canada Line. This resulted in a final map that only showed about half of Vancouver. These missing values were cropped out of the weighted sum analysis and often led to improper data sets. As a result, I had to merge this later to the major roads network.
Euclidean Distance Function
The Euclidean distance function creates a raster layer that displays the proximity to a certain feature. Each cell is evaluated as the overall distance between the particular cell and the nearest location of the feature in question in meters. This tool was used to identify areas that are within a certain range of a particular trait or property. Euclidean distance also created surface layers that could easily be used in the weighted sum Function.
Future Studies
In follow up studies, it is important to focus on each of the four bird categories and to investigate particular species. This will allow researchers to better tailor the weighting schemes. Having more in-depth approaches to specific species can allow us to incorporate factors like elevation and temperature. Additionally, it will be crucial to expand the study over a broader geographical range such as the entire Metro Vancouver region. This will provide better insight into how each species is responding: for example, the eagle population has actually skyrocketed outside of Vancouver. While we didn’t see this in my initial project findings, they have become common sightings. It would be insightful to map how birds are being spatially distributed alongside their overall trends.