J Plant Ecol ›› 2016, Vol. 9 ›› Issue (6): 742-751 .DOI: 10.1093/jpe/rtw009

• Research Articles • Previous Articles     Next Articles

Modeling habitat distribution of Cornus officinalis with Maxent modeling and fuzzy logics in China

Bo Cao1,?, Chengke Bai1,2,*,?, Linlin Zhang1, Guishuang Li1 and Mingce Mao3   

  1. 1 College of Life Sciences, Shaanxi Normal University, Xi'an 710119, China; 2 Co-Innovation Center for Qinba Regions' Sustainable Development, Shaanxi Normal University, Xi'an 710119, China; 3 Shaanxi Climate Center, Meteorological Bureau of Shaanxi Province, Xi'an 710064, China
  • Received:2015-05-25 Accepted:2016-02-02 Published:2016-12-02
  • Contact: Bai, Chengke

Modeling habitat distribution of Cornus officinalis with Maxent modeling and fuzzy logics in China

Abstract: Aims Predicting suitable habitat distribution is an effective way to protect rare or endangered medicinal plants. Cornus officinalis is a perennial tree growing in forest edge and its air-dried pericarp is one of the traditional Chinese medicines (TCM) with significant medicinal values. In recent years, C. officinalis has undergone severe degeneration of its natural habitat owing to growing market demands and unprecedented damage to the forests. Moreover, the degeneration of suitable habitat has threatened the supply of medicinal materials, and even led to the extinction of some engendered medicinal plant species. In this case, there is a great risk to introduce and cultivate medicinal plants if planners determine the suitable cultivation regions based on personal subjective experience alone. Therefore, predicting suitable potential habitat distribution of medicinal plants (e.g. C. officinalis) and revealing the environmental factors determining such distribution patterns are important to habitat conservation and environmental restoration.
Methods In this article, we report the results of a study on the habitat distribution of C. officinalis using maximum entropy (Maxent) modeling and fuzzy logics together with loganin content and environmental variables. The localities of 106 C. officinalis in China were collected by our group and other researchers and used as occurrence data. The loganin content of 234 C. officinalis germplasm resources were tested by high-performance liquid chromatography (HPLC) and used as content data. 79 environmental variables were selected and processed with multicollinearity test by using Pearson Correlation Coefficient (r) to determine a set of independent variables. The chosen variables were then processed in the fuzzy linear model according to the cell values (maximum, minimum) of localities with estimated loganin content. The SDMtoolbox was used to spatially rarefy occurrence data and prepare bias files. Furthermore, combined Maxent modeling and fuzzy logics were used to predict the suitable habitat of C. officinalis. The modeling result was validated using null-model method.
Important findings As a result, six environmental factors including tmin3, prec3, bio4, alt, bio12 and bio3 were determined as key influential factors that mostly affected both the habitat suitability and active ingredient of C. officinalis. The highly suitable regions of C. officinalis mainly distribute in a 'core distribution zone' of the east-central China. The statistically significant AUC value indicated that combined Maxent modeling and fuzzy logics could be used to predict the suitable habitat distribution of medicinal plants. Furthermore, our results confirm that ecological factors played critical roles in assessing suitable geographical regions as well as active ingredient of plants, highlighting the need for effective habitat rehabilitation and resource conservation.

Key words: Cornus officinalis, habitat distribution, Maxent modeling, fuzzy logics, medicinal plant

摘要:
Aims Predicting suitable habitat distribution is an effective way to protect rare or endangered medicinal plants. Cornus officinalis is a perennial tree growing in forest edge and its air-dried pericarp is one of the traditional Chinese medicines (TCM) with significant medicinal values. In recent years, C. officinalis has undergone severe degeneration of its natural habitat owing to growing market demands and unprecedented damage to the forests. Moreover, the degeneration of suitable habitat has threatened the supply of medicinal materials, and even led to the extinction of some engendered medicinal plant species. In this case, there is a great risk to introduce and cultivate medicinal plants if planners determine the suitable cultivation regions based on personal subjective experience alone. Therefore, predicting suitable potential habitat distribution of medicinal plants (e.g. C. officinalis) and revealing the environmental factors determining such distribution patterns are important to habitat conservation and environmental restoration.
Methods In this article, we report the results of a study on the habitat distribution of C. officinalis using maximum entropy (Maxent) modeling and fuzzy logics together with loganin content and environmental variables. The localities of 106 C. officinalis in China were collected by our group and other researchers and used as occurrence data. The loganin content of 234 C. officinalis germplasm resources were tested by high-performance liquid chromatography (HPLC) and used as content data. 79 environmental variables were selected and processed with multicollinearity test by using Pearson Correlation Coefficient (r) to determine a set of independent variables. The chosen variables were then processed in the fuzzy linear model according to the cell values (maximum, minimum) of localities with estimated loganin content. The SDMtoolbox was used to spatially rarefy occurrence data and prepare bias files. Furthermore, combined Maxent modeling and fuzzy logics were used to predict the suitable habitat of C. officinalis. The modeling result was validated using null-model method.
Important findings As a result, six environmental factors including tmin3, prec3, bio4, alt, bio12 and bio3 were determined as key influential factors that mostly affected both the habitat suitability and active ingredient of C. officinalis. The highly suitable regions of C. officinalis mainly distribute in a 'core distribution zone' of the east-central China. The statistically significant AUC value indicated that combined Maxent modeling and fuzzy logics could be used to predict the suitable habitat distribution of medicinal plants. Furthermore, our results confirm that ecological factors played critical roles in assessing suitable geographical regions as well as active ingredient of plants, highlighting the need for effective habitat rehabilitation and resource conservation.