| In recent years, a group of small-bodied fishes with pelagic eggs in United States were near extinction. In this paper, federally endangered small-bodied cyprinids (Shiner in Arkansas River, Notropis girardi, "ARS" for short) were studied. The study aimed at providing valuable references for the protection of river eco-system and endangered fish. The study was funded by United States Geological Survey program "Biotic and Abiotic Factors Related to Declining Distribution of a Federally Threatened Cyprinid". Specific topics of this study are as follows:(1) The study of methods for improving niche model transferabilityWe took MAXENT niche model as an example to define niche model transferability based on the concept of "projection" and Parolo’s theory about calculating Spearman’s rho correlation coefficient for the predicted distributions before and after model’s transformation. We presented four methods to improve the transferability by reducing the model dimension and complexity:variables selection based on the Spearman’s rho correlation coefficient|r| of the environment variables, variables selection based on the variance inflation factor|VIF|, variables reorganization based on principal component analysis PCA and adjusting the parameter "Regularization Multiplier" of the model. We compared the improvement of model transferability by using the "backcasting" and "forecasting" rho based on the four methods. The results showed that the four methods could effectively increase the "backcasting" and "forecasting" transferability of the model. Improving "Regularization Mutiplier" value could help improve the model transferability significantly. But the high "Regularization Mutiplier" value could make the response curves of the relationship between the species distribution probability and environment variable too smooth and cause the information loss. We proposed to take the model fit into consideration and combine variable correlation analysis with model parameter adjustment for improving the transferability of the model. This study could provide reference for niche model accuracy validation and improvement of niche model transferability.(2) The study of ARS habitat suitability analysis methodBased on geographic information system (GIS) and remote sensing (RS) technology, we collected natural environment data that may affect the potential distribution of ARS, including:biological climatic data, land cover data, geological data, slope and height data, river flow regime data and invasive species data. Also, we calculated factors related to human activities, such as point sources data near the main rivers, groundwater exploitation data and dams distribution data. To analyze the impact of natural environment and human activities on the potential distribution of ARS, we modelled the distribution of ARS in two different periods based on MAXENT. The results showed that the contribution rate of the river flow regime (e.g.,the average annual flow and river levels) and human activities (e.g., groundwater extraction factor and point sources factor) were greater than 5%, which means that these factors have affected the distribution of potential habitat ARS most. The response curve shape of the relationship between ARS predicted distribution and each factor were different in two periods. The main reason was that the distribution quantity and range have changed in two periods. By modelling the potential ARS habitat within a large-scale watershed and a long time period, the study could help determine the specific suitable living environment for the endangered fish.(3) The study of identification of eco-hydrological factors related to ARS spawningIn this study, daily flow data of 40 flow gages distributed in Arkansas River watershed from the year 1950 to 2010 were downloaded. According to the hydrological requirements for ARS spawning, we calculated three important ecological hydrological factors:high flow, high flow duration and zero flow days, In combination with other environmental factors, three species distribution models based on the theory of maximum entropy (MAXENT) were established to analyze the natural flow regime change before and after the construction of large hydraulic engineering. The results showed that, of the three flow metrics, high flow made the highest contribution to the model fit (historic 65.8%, current 55.7%) and had the most significant influence on the fish spawning. In historic period, the range of the flow metric when the species occurrence probability were predicted high could provide valuable references for reconstruction and operation of river eco-system.(4) The study of ARS potential habitat change analysis under climate changeThe latest climate scenarios data released by IPCC AR5 were downloaded. Based on the ecological meaning of the climatic factors and the correlation among the factors, we chose five climatic factors to simulate ARS potential suitable habitat change based on the CCSM 4.0 community climate system model in 2030 and 2050 to 2070. As the climate situation of the Arkansas River watershed were different in western and eastern sub region, we proposed to divide the whole study area into two sub region based on the minimum envelope rectangle. By comparing the Gain rate and Loss rate of the sub region and overall region, the accuracy of the model was verified. The results showed that the range of Gain rate and Loss rate were within 1±0.3, which proved that the predicted results of sub region models and whole region model had consistency and the model results were accurate. By mapping the change trend of the ARS distribution pattern, the study indicated that ARS had the trend to move to the upper reaches of the river watershed under the climate change. In eastern sub-region, the ARS distribution probability for most river segments were decreasing, except for a few segments in the intersection of Cimmaron and Arkansas River. The method could be helpful to reduce the impacts of future climate change on the endangered species. |
PDF Full Text Request