2025-08-03
Today is about urban green space, and we learn that people really like plants and birds.
Papers
This week I have mainly read about urban biodiversity to understand how it can be improved. I’d like to present 2 papers on the topics.
Knowing that when urban biodiversity is considered, it has often been often ad-hoc or reactive, it is interesting to see what matters, and how (or if) machine learning can help.
The first paper is by Soanes et al’s Conserving urban biodiversity: Current practice, barriers, and enablers, Conservation Letters 16 (2023). This paper presents an interesting study on how urban environmental manager can and have improved urban biodiviersity. One interesting take-away is that, while human-nature spaces are frequently highlighted as a motivation for urban biodiversity, designing for those spaces is rare (5% of all actions describes by the managers). Another interesting point is that one of the barrier is simply an apathy regarding biodiversity loss. Especially they give example of projects that failed because of citizens opposition due to a lack of outreach and participation possibilities. This study is interesting because it highlights some key component of success in urban biodiversity conservation and the misalignment between our priorities and what is really happening on the ground.
The meta-analysis by Rega-Brodsky et al Urban biodiversity: State of the science and future directions (Urban Ecosystems 25 ,2022) showcase the limit of the study on urban biodiversity conservation. Indeed, those have been mostly ignored the global south, even though most biodiversity can be found there. Furthermore, even though nature-based solutions are becoming more popular, at the time of writing, only 5% of the studies linked biodiversity and ecological functions.
An interesting point is that, overall, birds and plants are over-represented in studies while other taxa are wildly under-represented with less than 20 publications per years.
In their study, they also point to the lack of long-term monitoring to understand relationships and temporal trends, and show a lack of diversity in the type of landscape studied, with an over-representation of forests, managed public land, and private yards.
On the other hand, rudeal vegetation, costal dunes, and deserts are under-represented.
Finally, the impact of restoration/management was also seldom studied (8.5%), and again mostly on plant and birds in forests.
I believe that machine learning, sensor fusion, and robotics can be an interesting avenue to provide new solutions for those challenges.
As a side note, they point to the journals where most work were published so that if one want to explore more on this subject, one should look into Urban Ecosystems (13%), Landscape and Urban Planning (11%), Urban Forestry & Urban Greening (5%), and Biological Conservation (4%).