Assessment of American Bullfrog (Lithobates catesbeianus) spreading in the Republic of Korea using rule learning of elementary cellular automata

Abstract

Abstract The spread of American Bullfrog, one of the 100 of the World’s Worst Invasive Alien Species, has a great impact on the surrounding ecosystem. Little is known about the tendancy and pattern of how they are spreading in South Korea geographically. It is important to study the tendancy of their spreading so that a proper mitigation can be applied when needed. This study is based on the results of national surveys that observed the distribution. The entire data is divided into 25 regional clusters using the divisive hierarchical clustering method. In order to estimate the degree of spreading, a sequence of spatial distribution is constructed for each cluster using the agglomerative clustering method. ECA(elementary cellular automata) is introduced to find rules governing the pattern variation in the sequence. Each cell represents either the observed or unobserved site of bullfrog. The number of Bullfrog Observed Site (BOS) in a sequence of each cluster is counted and used to define the spreading intensity. The rules of ECA are trained and learned by the CNN(Convolution Neural Network) method and used to estimate and predict the spreading intensity by counting the number of BOS. Taking environmental factors into account, habitat suitability is used and obtained using Maxent. The spreading intensity is multiplied by the habitat suitability to get an assessment of bullfrogs spreading. The relative spreading assessment is estimated, which is classified into 4 groups; spreading intensively, spreading slowly, maintaining or declining population.