Bird Habitat Suitability Analysis and Digital Twin Model

Urban green spaces are closely related to the abundance and biodiversity of birds by providing important habitats and together contribute to ecosystem health. This project aims to guide the University of British Columbia Botanical Garden to create Bird-friendly green spaces by using LiDAR data to analyze and map UBCBG's bird habitat suitability and create a 3D digital twin model of UBCBG in the open source game engine Minetest to increase 3D visualization and aid in landscape planning. By extracting the Canopy Height Model (CHM) using LiDAR data and performing individual tree segmentation, the derived metrics were used to identify trees with the highest bird habitat suitability index. The results showed that the suitability index ranges from -0.0016 to 0.5187, with a mean value of 0.2051. There are 68 trees with high suitability above the 0.4 intervals which have significance to bird populations and are worthy of being protected, accounting for only 3.38% of the total trees. They usually have a low vertical complexity index and foliage height diversity but are characterized by very tall trees with relatively large tree crowns. The Digital Elevation Model (DEM), Canopy Height Model (CHM) generated by LiDAR data were visualized in Minetest's UBCBG's 3D digital twin model using real terrain mod as topography and vegetation layers, while bird habitat suitability was used to symbolize the tree canopy layer. This study is highly relevant for landscape adaptation and planning in conjunction with other management considerations to support bird-friendly green spaces. The digital twin model can be used for educational and promotional purposes, and for landscape planning and aesthetic design with the consideration of bird conservation.

Keywords: LiDAR, GIS, Minetest, digital twin, habitat, landscape planning, bird ecology

Birds play an essential role in the ecosystem by controlling the ecological balance, spreading seeds and pollinating, serving as indicators of ecosystem health, and providing aesthetic, cultural, and economic value for human beings (Tabur, et a., 2010). However, the diversity and abundance of bird species are at risk worldwide in the face of climate change and human-induced land-use change (UBC Botanical Garden, 2018). Habitat loss is the primary cause that could lead to severe population declines in bird populations (Dolman & Sutherland, 1995). Conversely, bird diversity and abundance can be facilitated by providing a habitat with particular habitat features (e.g., mature trees and fruit-bearing plants) (Melles et al., 2003).  The Guidelines: Bird Friendly Design - Explanatory Note (City of Vancouver, 2015)  combines multiple studies and the specifics of native bird species in Vancouver to provide useful advice on creating bird-friendly landscapes designed to enhance the overall abundance and diversity of birds in Vancouver. It proposed that landscape features such as patch size, tree species, habitat type, vertical vegetation structure, plant diversity, etc. are key factors in measuring the quality of the landscape as a habitat for native birds.

These landscape features, which are closely related to birds, can be monitored and assessed using remote sensing techniques (Duro et al., 2007). Compared to passive remote sensing techniques which have been widely used to derive habitat variables (Graf et al., 2009), LiDAR (Light Detection And Ranging) data is laser scanning obtained by remote sensing which can provide the extra structural information at the landscape scale that is crucial for not only birds but also overall wildlife (Graf et al., 2009, and Tattoni et al., 2012). LiDAR data can demonstrate a more realistic spatial pattern of landscape and the derived variables can be interpreted more directly from an ecological perspective and in habitat management (Tattoni et al., 2012). Therefore, the habitat suitability analysis can be conducted by using Lidar data and derived metrics can be used to evaluate landscape features related to the abundance and biodiversity of birds.

There is a growing trend of integrating geospatial data and gaming engines to create a digital twin model of the real-world landscape in video game engines for the purposes of business, education, landscape planning, etc.(Tauscher & Le, 2021). Minetest is an open-source voxel game that allows the use of lua (a programming language) API (application programming interface) to create diverse customized voxel games (Minetest, 2022). There are now an increased number of cases where real-world raster and vector geographic databases are imported into Minetest to create real-world 3D voxel game versions (e.g., Lecordix et al., 2019) for educational, commercial, research, and other diverse uses. Such attempts not only visualize the presentation of realistic geographic maps in a 3D voxel game perspective, but also are interactive, and entertaining for the players, allowing for remote work and communication, and allowing for also engaging the younger generation for example in geospatial learning and management (Tauscher& Le, 2021). 

This project aims to help the University of British Columbia Botanical Garden (UBCBG) to become a more bird-friendly green space by conducting a bird habitat suitability analysis and creating a digital twin model of UBCBG in Minetest to visualize the landscape and demonstrate the habitat suitability. The habitat suitability analysis is conducted based on the bird-friendly landscape design guidelines (City of Vancouver, 2015), in terms of vertical vegetation structure, human disturbance, and water accessibility. Moreover, the 3D voxel digital twin of UBCBG will be created in Minetest with the presentation of bird habitat suitability results, which will not only help UBCBG to conduct planning simulations, but also can be the tool for students to learn about tree and plant species, landscape management, etc., and the wider public to take a virtual tour of the garden. This will be an interesting and novel attempt to present realistic geospatial data through 3D voxel games, combine research and education in an entertaining and interactive way, and can also contribute to UBCBG's landscape planning and commercial promotion.

The study site is the University of British Columbia Botanical Garden (UBCBG) (Figure 1.) which is located in the southwestern corner of Vancouver, Canada with the coordinates of 49° 15' 15.588'' N 123° 15' 2.8224''. UBCBG was established in 1916 on the University of British Columbia Vancouver (UBCV) campus, it collects and displays ~30,000 plants from ~8000 accessions (UBC Botanical Garden, 2022). The UBCBG supports a wide variety of biodiversity species, with the University of British Columbia Botanical Garden (UBCBG) being particularly valuable as a habitat for a variety of bird species. UBCBG has diverse and rich plant compositions that can provide adequate food resources for birds (e.g., seeds, leaves, flower nectar, pollen, etc.) and provide a habitat for breeding and protection (UBC Botanical Garden, 2018). In the past five years, more than 90 bird species were observed in UBCBG ( eBird  Canada, 2023) which indicates the attractiveness of UBCBG for birds. The UBCBG recognizes the importance of bird conservation and protection and hopes to maintain bird richness and biodiversity in the UBCBG by creating bird-friendly green spaces. and biodiversity.

A total of 2012 trees located at UBCBG were identified by segmenting Canopy Height Model(CHM) derived by Lidar point cloud, each with an individual Tree ID, on which metrics for bird habitat suitability analysis (i.e., tree height, crown size, distance to water, distance to human disturbance, VCI), latitude and longitude position information and suitability index were recorded. The suitability index ranges from -0.0016 to 0.5187, with a mean value of 0.2052, a median value of 0.2097, a variance of 0.0122, and a standard deviation of 0.1107.

All but one of the trees has a positive suitability index, of which 993 trees (49.35%) are in the moderate suitability range of 0.2000-0.4000 (Figure 7.), and there are 68 trees with high suitability above the 0.4000 intervals, accounting for only 3.38% of the total ratio. See the dark green section of Figure 6. These trees are less likely to appear clustered and are distributed throughout the UBCBG, but overall, more of the high-suitability trees are located in the middle and northeast corners of the UBCBG, in areas away from buildings and where the canopy is denser.

By using real terrain mod in Minetest to create a 3D digital twin model of UBCBG, including real-world terrain and individual trees, the model is able to simulate the actual landscape of UBCBG. The terrain is generated by the Digital Elevation Model (DEM), which realistically depicts the undulating changes in the UBCBG terrain (Figure 8). The individual trees in UBCBG are generated by the Canopy Height Model (CHM) (Figure 9.), which shows the distribution of trees in UBCBG, the height and crown size of individual trees, and the vertical vegetation structure they form together, helping to understand the complexity and variability of the overall habitat's vertical vegetation structure. As shown in Figure 10., the landscape including the terrain and individual trees is clearly visible, making this Minetest digital twin an excellent model for simulating bird habitats in UBCBG.

The bird suitability analysis in UBCBG conducted during this project has a strong reliance on Lidar-derived metrics, and Lidar data was also the primary source of data for subsequent 3D visualization of the UBC Botanical Garden within the Minetest Game Engine. However, other studies have shown that additional metrics from high-resolution RS products (land cover maps, country-wide ALS, Sentinel-1, and Sentinel-2) might provide additional habitat modeling information and the combination of Lidar and Sentinel can be a better predictor of habitat suitability (Koma et al., 2022), so it is expected that more accurate habitat suitability models for birds in the UBC Botanical Garden will be developed in the future using a diverse combination of remote sensing data and other data types.

Trees with the highest bird habitat suitability index generally had a low VCI (Vertical Complexity Index), but were characterized by very tall trees with relatively large crowns, meaning that from the data perspective, these large trees had generally low vertical vegetation structure complexity. This is due to the fact that the assessment of understorey vegetation by LiDAR is often less accurate under dense canopies, especially as the proportion of laser pulses reaching the lower forest strata decreases, and therefore the separate mapping of understorey vegetation could be considered (Martinuzzi et al., 2019) for more accurate and effective habitat assessment and habitat suitability model building. Another reason might be that these taller trees have less understory and layers under the canopy in UBC Botanical Garden. First, forests with open canopies tend to support more shrubs than forests with closed canopies (Bartemucci et al., 2006), and areas with high-suitability trees in the UBC Botanical Garden also have denser canopies and therefore may have fewer shrub layers. Another reason for this is the purposeful planning of the Botanical Garden and the design and aesthetics of the garden displays.

This bird suitability analysis was not validated and tested for accuracy because there was a lack of sufficient quantity and quality of independent data on the distribution or abundance of birds in the study area to validate the habitat suitability results. However, there is one citizen science project, eBird, that has recorded observations of birds, including information on the time of recording, the area, and the species of birds. But the amount of data available from this source is generally limited, the number of observations is generally low, and most of the data lack precise location information (although over 500 observations and 121 bird species have been recorded at UBCBG, only two observations have precise locations), so the amount of data is too small to support the analysis and validation of the suitability analysis results.

Overall, this study provides a map of bird habitat suitability for the UBC Botanical Garden, and the results of the bird habitat suitability analysis provide good insights and recommendations for the creation of Bird-friendly green space at the UBC Botanical Garden. This study is aimed at the overall abundance and diversity of birds at the study area level, but new analyses and models can be developed by changing or adding preferred habitat conditions if specific bird species are desired.

This project is a very good example that shows geospatial data is a good and useful data source to build a 3D digital twin model in an open-source game engine. By processing LiDAR point cloud, generating the Digital Elevation Model (DEM) and Canopy Height Model (CHM), which can be used to simulate the real terrain and vegetation of UBCBG in the Minetest game engine, allowing for 3D landscape visualization. In addition to visualization, more projects and individuals have developed more customized functions and game mods, such as adding tools like compasses, and calculators, or using them as teaching platforms, landscape planning, etc. We can expect in the future, this UBCBG digital twin model can be further developed as a virtual tour product of the UBC Botanical Garden with more development, open to students and visitors as well as the general public, for both educational and promotional purposes. It can also support the landscape planning and aesthetic design of the UBC Botanical Garden by visually changing the landscape and adding or subtracting objects. To better achieve these goals and support more uses, future directions for the project could include more development of the model, such as adding more tools and functionality to improve usability and interactivity.

Based on the bird habitat suitability analysis, it is evident that the University of British Columbia Botanical Garden provides a suitable habitat for birds. The analysis shows that most of the trees in the garden have a positive suitability index, and 49.35% of them are in the moderate suitability range, while 3.38% have high suitability for birds.

However, the bird habitat suitability analysis relies heavily on Lidar-derived metrics, which may not accurately assess understorey vegetation. Therefore, further studies combining Lidar and other remote sensing data sources could provide more accurate habitat modelling information. Also, the bird suitability analysis was not validated due to the lack of independent data on the distribution or abundance of birds in the study area.

Overall, the University of British Columbia Botanical Garden could be considered suitable habitat for birds, and future studies should consider additional data sources to improve habitat suitability models. It is also essential to continue monitoring and recording observations of birds in the garden to validate the habitat suitability results and ensure that conservation efforts remain effective.

Bradbury, R. B., Hill, R. A., Mason, D. C., Hinsley, S. A., Wilson, J. D., Balzter, H., ... & Bellamy, P. E. (2005). Modelling relationships between birds and vegetation structure using airborne LiDAR data: a review with case studies from agricultural and woodland environments. Ibis, 147(3), 443-452.

Carrasco, L., Giam, X., Papeş, M., & Sheldon, K. S. (2019). Metrics of lidar-derived 3D vegetation structure reveal contrasting effects of horizontal and vertical forest heterogeneity on bird species richness. Remote Sensing, 11(7), 743.

City of Vancouver (January 2015). Bird Friendly Design Guidelines Explanatory Note.  https://guidelines.vancouver.ca/guidelines-bird-friendly-design-explanatory-note.pdf  

Coops, N. C., Tompaski, P., Nijland, W., Rickbeil, G. J., Nielsen, S. E., Bater, C. W., & Stadt, J. J. (2016). A forest structure habitat index based on airborne laser scanning data. Ecological Indicators, 67, 346-357

Dalponte, M., & Coomes, D. A. (2016). Tree‐centric mapping of forest carbon density from airborne laser scanning and hyperspectral data. Methods in ecology and evolution, 7(10), 1236-1245.

Dolman, P. M., & Sutherland, W. J. (1995). The response of bird populations to habitat loss. Ibis, 137, S38-S46.

Douss, R., & Farah, I. (2022). Extraction of individual trees based on Canopy Height Model to monitor the state of the forest. Trees, Forests And People, 8, 100257. doi: 10.1016/j.tfp.2022.100257

Erfanifard, Y., Stereńczak, K., Kraszewski, B., & Kamińska, A. (2018). Development of a robust canopy height model derived from ALS point clouds for predicting individual crown attributes at the species level. International Journal of Remote Sensing, 39(23), 9206-9227.

Graf, R. F., Mathys, L., & Bollmann, K. (2009). Habitat assessment for forest dwelling species using LiDAR remote sensing: Capercaillie in the Alps. Forest Ecology and Management, 257(1), 160-167.

Hill, R. A., & Hinsley, S. A. (2015). Airborne lidar for woodland habitat quality monitoring: Exploring the significance of lidar data characteristics when modelling organism-habitat relationships. Remote Sensing, 7(4), 3446-3466.

Hyyppä, J., Hyyppä, H., Leckie, D., Gougeon, F., Yu, X., & Maltamo, M. (2008). Review of methods of small‐footprint airborne laser scanning for extracting forest inventory data in boreal forests. International Journal of Remote Sensing, 29(5), 1339-1366.

Jean-Romain Roussel, Tristan R.H. Goodbody, Piotr Tompalski. (2022-08-17). The lidR package.  https://r-lidar.github.io/lidRbook/  

Khosravipour, A., Skidmore, A. K., Isenburg, M., Wang, T., & Hussin, Y. A. (2014). Generating pit-free canopy height models from airborne lidar. Photogrammetric Engineering & Remote Sensing, 80(9), 863-872.

Lecordix, F., Kumarasamy, S., Da Graca, F., & Kriat, S. (2019, July). Preparation of the Future with Minecraft® on Demand. In Proceedings of the ICA (Vol. 2, pp. NA-NA). Copernicus GmbH.

Liu, H., & Dong, P. (2014). A new method for generating canopy height models from discrete-return LiDAR point clouds. Remote sensing letters, 5(6), 575-582.

Mammides, C., Kadis, C., & Coulson, T. (2015). The effects of road networks and habitat heterogeneity on the species richness of birds in Natura 2000 sites in Cyprus. Landscape Ecology, 30(1), 67-75.

Melles, S., Glenn, S., & Martin, K. (2003). Urban bird diversity and landscape complexity: species–environment associations along a multiscale habitat gradient. Conservation Ecology, 7(1).

Minetest - Open source voxel game engine. (2022). Retrieved 8 December 2022, from https://www.minetest.net/

Müller, J., Stadler, J., & Brandl, R. (2010). Composition versus physiognomy of vegetation as predictors of bird assemblages: The role of lidar. Remote Sensing of Environment, 114(3), 490-495.

Tattoni, C., Rizzolli, F., & Pedrini, P. (2012). Can LiDAR data improve bird habitat suitability models?. Ecological Modelling, 245, 103-110.

Tauscher, H., & Le, V. T. (2021). MAP GENERATION FOR RETRO GAMING ENGINES. ISPRS Annals of Photogrammetry, Remote Sensing & Spatial Information Sciences.

UBC Botanical Garden. (2022). Plant Collections - Retrieved 29 October 2022, from  https://botanicalgarden.ubc.ca/research-collections/plant-collections/ 

UBC Botanical Garden. (2018). UBCBG bird brochure 2018.  https://botanicalgarden2015.sites.olt.ubc.ca/files/2018/10/ubcbg-adult-ed-bird-brochure-2018.pdf 

University of British Columbia. Campus and Community Planning. (2021). [University of British Columbia Point Grey Campus Lidar], 2021 (University of British Columbia. Campus and Community Planning).  Abacus Data Network, V1.  https://abacus.library.ubc.ca/dataset.xhtml?persistentId=hdl:11272.1/AB2/Y5KQNB  

University of British Columbia. Campus and Community Planning. (2016). UBCGeodata, [November 2016] (University of British Columbia. Campus and Community Planning). Abacus Data Network, V2.  https://abacus.library.ubc.ca/dataset.xhtml?persistentId=hdl:11272.1/AB2/S15BIR  

Zhang, W., & Montgomery, D. R. (1994). Digital elevation model grid size, landscape representation, and hydrologic simulations. Water resources research, 30(4), 1019-1028.

Koma, Z., Seijmonsbergen, A. C., Grootes, M. W., Nattino, F., Groot, J., Sierdsema, H., ... & Kissling, W. D. (2022). Better together? Assessing different remote sensing products for predicting habitat suitability of wetland birds. Diversity and distributions, 28(4), 685-699.

Bakx, T. R. M., Koma, Z., Seijmonsbergen, A. C., & Kissling, W. D. (2019). Use and categorization of light detection and ranging vegetation metrics in avian diversity and species distribution research. Diversity and Distributions, 25(7), 1045–1059. https://doi.org/10.1111/ddi.12915

Martinuzzi, S., Vierling, L., Gould, W., Falkowski, M., Evans, J., Hudak, A., & Vierling, K. (2009). Mapping snags and understory shrubs for a LiDAR-based assessment of wildlife habitat suitability. Remote Sensing Of Environment, 113(12), 2533-2546. doi: 10.1016/j.rse.2009.07.002

Bartemucci, P., Messier, C., & Canham, C. D. (2006). Overstory influences on light attenuation patterns and understory plant community diversity and composition in southern boreal forests of Quebec. Canadian Journal of Forest Research, 36(9), 2065-2079.