Calling Phenology of Two Frog Species in South Korean Rice Paddies Using Automated Call Detection

TL;DR

This study tracks frog calling patterns in South Korean rice paddies over five years to understand how they respond to weather and farming.

Contribution

The study provides new insights into species-specific calling phenology and the effectiveness of automated acoustic monitoring in agroecosystems.

Findings

Dryophytes japonicus has a short, synchronized calling peak in early summer, strongly influenced by temperature.

Pelophylax nigromaculatus calls earlier and longer, with activity affected by both temperature and humidity.

Automated acoustic monitoring proved effective for long-term biodiversity observation in agricultural wetlands.

Abstract

Amphibians are sensitive to changes in their surroundings, and the phenology of breeding calls can reflect how they respond to weather and farming practices. We studied two common frog species in South Korean rice paddies over five years using sound recorders. This made it possible to track calling activity over long periods without disturbing the animals. Dryophytes japonicus showed a short and synchronized calling peak in early summer, while Pelophylax nigromaculatus began earlier in spring and continued calling for longer, although at lower levels. Temperature was the main factor driving calling in both species, with the D. japonicus showing especially strong responses. Humidity also played a role, particularly for the P. nigromaculatus, which called more actively when conditions were moist. These results reveal how frogs adapt their reproductive behavior to agricultural wetland environments and changing weather conditions. Our study highlights the ecological value of rice paddies as breeding habitats and shows that automated acoustic monitoring can be a powerful tool for long-term biodiversity observation and conservation planning.

Amphibian breeding phenology provides key insights into species’ sensitivity to climatic and anthropogenic drivers. We used passive acoustic monitoring (PAM) with automated call detection to examine the calling activity of Dryophytes japonicus and Pelophylax nigromaculatus in South Korean rice paddies across five breeding seasons (2018–2022). Both species exhibited distinct seasonal patterns: D. japonicus showed a synchronous and concentrated calling peak in mid-June (GAM deviance explained = 34%), whereas P. nigromaculatus initiated calling earlier and maintained a longer, less synchronized calling period extending into July (GAM deviance explained = 19%). Zero-inflated negative binomial models demonstrated that temperature was the strongest predictor of calling activity in both species, though responses to humidity and wind differed. D. japonicus maintained high calling rate under warm conditions, with only modest suppression at high humidity, whereas P. nigromaculatus was strongly inhibited by combined warm and humid conditions. These results establish a detailed information on the calling phenology of D. japonicus and P. nigromaculatus in East Asian agroecosystems highlight species-specific sensitivities to local weather variables. Our findings demonstrate that automated acoustic monitoring offers an efficient way to document ecological responses to weather variability and may serve as a long-term tool to track phenological shifts under climate change. Future advances in sound analysis, including the integration of deep-learning algorithms and cross-species detection frameworks, could further improve automated biodiversity monitoring in complex agricultural landscapes.

Figures