How are bats monitored? Exploring fixed and multimodal sensing techniques

Bats fly at night, but scientists don’t need a flashlight to study them. Fixed sensors – infrared, cameras, radar and microphones – can gather data on thousands of individuals without disturbing them. Infrared detects body heat, creating a thermal map of flight. High‑speed cameras capture rapid movements, while radar tracks trajectories over long distances. Microphones, on the other hand, listen to echolocation calls, turning sounds into activity maps. Each method has strengths: infrared works in total darkness, cameras show visual detail, radar covers large areas and acoustics identify species by their unique calls. Yet none is perfect: cameras need enough light, radar can confuse small insects and microphones pick up ambient noise. Researchers combine several sensors into multimodal systems, gaining a fuller picture of flight behavior, migration routes and responses to climate change. These data help us understand how bats move, which habitats they prefer and how to protect threatened colonies. In the future, smaller sensors, AI for real‑time analysis and distributed monitoring networks promise even more effective and eco‑friendly bat surveillance.

Bats belong to the order Chiroptera, characterized by membranous wings supported by elongated bones. Their ecology is tightly linked to echolocation: they emit ultrasonic pulses (20–200 kHz) and analyze the echo to locate insects and obstacles. This ability is exploited in acoustic sensors, where machine‑learning algorithms classify calls based on spectral parameters (fundamental frequency, duration, modulation). X‑band radar (8–12 GHz) provides angular resolution of a few degrees, sufficient to distinguish bat groups in flight. However, the radar cross‑section of small mammals is low, requiring temporal integration to improve signal‑to‑noise ratio. Thermal cameras with 640×480 px resolution and microbolometer sensors can detect temperature differences below 0.1 °C, allowing tracking of individual bats even in urban settings. Multimodal integration uses data‑fusion techniques based on Kalman filters or convolutional neural networks, combining spatial information (radar, video) and temporal data (acoustics) to reconstruct 3‑D trajectories with average error under 0.5 m. Additionally, passive RFID tags powered by the radar’s electromagnetic field enable individual identification without invasive implants. These innovations pave the way for longitudinal studies on seasonal migrations, climate‑change impacts, and zoonotic pathogen spread, providing crucial data for bat conservation.