Ex-situ conservation can be a valuable tool to help threatened species boost natural populations with captive-born specimens or reintroduce locally extinct animals. Captive breeding can be particularly effective for small species with many offspring, such as animals with the r-breeding strategy. Furthermore, a controlled environment can also help us discover more about species' biology, especially those less studied and more challenging to track in nature. Despite the large number of species kept in captivity, many remain understudied. Often, there needs to be more information about their biology and the effects of management on their well-being. This knowledge gap is partly due to data collection's highly time-consuming and expensive nature, which is frequently avoided for non-iconic animals like cold-blooded vertebrates. In this study, I described and tested a cost-effective, easy-to-use, and automatic method to monitor the activity of reptiles in a changing captive environment. I investigated the effect of an increasing number of furnishings in social groups of Australian thick-tailed geckos (Underwoodisaurus milii, Bory de Saint-Vincent, 1823) using an infrared security camera and the software DeepLabCut to track animals in arenas for approximately 60 days. By calculating space use, proximity between individuals, and time budgets, I identified these metrics as valuable proxies for evaluating the effects of captive settings and management. My findings demonstrate that changes in environmental complexity significantly influenced gecko spatial behaviour, while automated tracking provided efficient, scalable insights into their activity patterns. This study underscores the scalability of inexpensive security cameras and machine learning tools for ex-situ research and welfare monitoring, with implications for improving the management and understanding of cryptic and understudied species.
Ex-situ conservation can be a valuable tool to help threatened species boost natural populations with captive-born specimens or reintroduce locally extinct animals. Captive breeding can be particularly effective for small species with many offspring, such as animals with the r-breeding strategy. Furthermore, a controlled environment can also help us discover more about species' biology, especially those less studied and more challenging to track in nature. Despite the large number of species kept in captivity, many remain understudied. Often, there needs to be more information about their biology and the effects of management on their well-being. This knowledge gap is partly due to data collection's highly time-consuming and expensive nature, which is frequently avoided for non-iconic animals like cold-blooded vertebrates. In this study, I described and tested a cost-effective, easy-to-use, and automatic method to monitor the activity of reptiles in a changing captive environment. I investigated the effect of an increasing number of furnishings in social groups of Australian thick-tailed geckos (Underwoodisaurus milii, Bory de Saint-Vincent, 1823) using an infrared security camera and the software DeepLabCut to track animals in arenas for approximately 60 days. By calculating space use, proximity between individuals, and time budgets, I identified these metrics as valuable proxies for evaluating the effects of captive settings and management. My findings demonstrate that changes in environmental complexity significantly influenced gecko spatial behaviour, while automated tracking provided efficient, scalable insights into their activity patterns. This study underscores the scalability of inexpensive security cameras and machine learning tools for ex-situ research and welfare monitoring, with implications for improving the management and understanding of cryptic and understudied species.
Automated behavioural data collection in a dynamic captive environment: using DeepLabCut with Underwoodisaurus milii (Bory De Saint-Vincent, 1823)
FERRERO, VITTORIO
2023/2024
Abstract
Ex-situ conservation can be a valuable tool to help threatened species boost natural populations with captive-born specimens or reintroduce locally extinct animals. Captive breeding can be particularly effective for small species with many offspring, such as animals with the r-breeding strategy. Furthermore, a controlled environment can also help us discover more about species' biology, especially those less studied and more challenging to track in nature. Despite the large number of species kept in captivity, many remain understudied. Often, there needs to be more information about their biology and the effects of management on their well-being. This knowledge gap is partly due to data collection's highly time-consuming and expensive nature, which is frequently avoided for non-iconic animals like cold-blooded vertebrates. In this study, I described and tested a cost-effective, easy-to-use, and automatic method to monitor the activity of reptiles in a changing captive environment. I investigated the effect of an increasing number of furnishings in social groups of Australian thick-tailed geckos (Underwoodisaurus milii, Bory de Saint-Vincent, 1823) using an infrared security camera and the software DeepLabCut to track animals in arenas for approximately 60 days. By calculating space use, proximity between individuals, and time budgets, I identified these metrics as valuable proxies for evaluating the effects of captive settings and management. My findings demonstrate that changes in environmental complexity significantly influenced gecko spatial behaviour, while automated tracking provided efficient, scalable insights into their activity patterns. This study underscores the scalability of inexpensive security cameras and machine learning tools for ex-situ research and welfare monitoring, with implications for improving the management and understanding of cryptic and understudied species.File | Dimensione | Formato | |
---|---|---|---|
TESI_Vittorio_Ferrero_ECAU2024_signed.pdf
non disponibili
Dimensione
20.92 MB
Formato
Adobe PDF
|
20.92 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14240/9702