野桂花和管花木犀的适宜分布区及主要气候变量分析
|
摘要
基于野桂花[Osmanthus yunnanensis(Franch.) P. S. Green]和管花木犀(O.delavayi Franch.)的分布记录和气候变量,利用MaxEnt模型预测其现代适宜分布区和未来(2070年)潜在适宜分布区,利用贡献率、置换重要值和Jackknife检验评价影响其分布的主要气候变量,并利用限制因子映射工具模拟各分布区主要气候变量的分布模型。结果表明:当正则化参数为0.5、参数组合为线性+二次型(L+Q)时,野桂花和管花木犀MaxEnt模型的AIC值最小,分别为1 170.4和817.9,AUC值最大,分别为0.976和0.948,表明优化参数后的MaxEnt模型预测精度非常高。利用该模型预测的野桂花和管花木犀的现代适宜分布区相似,主要包括四川、云南、广西西部、西藏东南部和贵州西部,其适生区面积分别占中国总面积的3.52%和4.21%,高度适生区面积分别占中国总面积的1.02%和1.14%。在RCP8.5气候情景下,野桂花的未来潜在适宜分布区向东部和北部扩张,适生区和高度适生区的面积分别为8.21%和0.30%;而管花木犀的未来潜在适宜分布区向西部和北部扩张,适生区和高度适生区的面积分别为4.41%和1.10%。依据贡献率、置换重要值和Jackknife检验,影响野桂花和管花木犀分布的主要气候变量均为气温季节性变化、年平均气温和年降水量;并且,在存活概率为0.5时,野桂花适宜分布区3个气候变量的方差范围大于管花木犀,说明野桂花的生态适应范围更广。另外,年平均气温限制了野桂花分布区的东界,年降水量制约其向北移动;年降水量限制了管花木犀分布区的南界。研究结果显示:优化参数后的MaxEnt模型预测结果非常准确,预测的野桂花和管花木犀的现代适宜分布区均为四川、云南、广西西部、西藏东南部和贵州西部,而其未来潜在适宜分布区却存在差异;温度和降水是影响其分布的主要气候变量,但不同分布区的主要气候变量却存在差异。
Based on distribution records and climatic variables of Osmanthus yunnanensis(Franch.) P. S. Green and O. delavayi Franch., their present suitable distribution areas and potential suitable distribution areas in the future(in 2070) were predicted by using MaxEnt model, and the main climatic variables affecting their distributions were evaluated by using contribution rate, permutation importance, and Jackknife test, meanwhile, the distribution models of main climatic variables in each distribution area were simulated by using limiting factor mapping tools. The results show that when regularization parameter of 0.5 and parameter combination of linearity+quadratic type(L+Q), AIC values of MaxEnt models of O. yunnanensis and O. delavayi are the smallest with values of 1 170.4 and 817.9, respectively, and AUC values are the largest with values of 0.976 and 0.948, respectively, indicating that the prediction accuracy of MaxEnt model after parameter optimization is very high. The present suitable distribution areas of O. yunnanensis and O. delavayi predicted by this model are similar, which mainly include Sichuan, Yunnan, West Guangxi, Southeast Tibet, and West Guizhou, and their areas of suitable distribution area account for 3.52% and 4.21% of total area of China, respectively, while their areas of highly suitable distribution area account for 1.02% and 1.14% of total area of China, respectively. Under the climatic scenario of RCP8.5, potential suitable distribution area in the future of O. yunnanensis expands to the east and north, and areas of suitable distribution and highly suitable distribution areas are 8.21% and 0.30%, respectively; while potential suitable distribution area in the future of O. delavayi expands to the west and north, and areas of suitable distribution and highly suitable distribution areas are 4.41% and 1.10%, respectively. According to contribution rate, permutation importance, and Jackknife test, main climatic variables affecting distributions of O. yunnanensis and O. delavayi are variation of temperature seasonality, annual mean temperature, and annual precipitation; in addition, variance ranges of three climatic variables of O. yunnanensis in suitable distribution area are greater than those of O. delavayi when survival rate of 0.5, indicating that the ecological adaptive range of O. yunnanensis is wider. Moreover, annual mean temperature constrains the east border of distribution area of O. yunnanensis, while annual precipitation constrains its movement toward north; annual precipitation does the south border of distribution area of O. delavayi. It is suggested that the prediction result of MaxEnt model after parameter optimization is very accurate, and the predicted present suitable distribution areas of O. yunnanensis and O. delavayi are Sichuan, Yunnan, West Guangxi, Southeast Tibet, and West Guizhou, but there is difference in their potential suitable distribution areas in the future; temperature and precipitation are the main climatic factors affecting their distributions, but there are differences in main climatic factors among different distribution areas.
Based on distribution records and climatic variables of Osmanthus yunnanensis(Franch.) P. S. Green and O. delavayi Franch., their present suitable distribution areas and potential suitable distribution areas in the future(in 2070) were predicted by using MaxEnt model, and the main climatic variables affecting their distributions were evaluated by using contribution rate, permutation importance, and Jackknife test, meanwhile, the distribution models of main climatic variables in each distribution area were simulated by using limiting factor mapping tools. The results show that when regularization parameter of 0.5 and parameter combination of linearity+quadratic type(L+Q), AIC values of MaxEnt models of O. yunnanensis and O. delavayi are the smallest with values of 1 170.4 and 817.9, respectively, and AUC values are the largest with values of 0.976 and 0.948, respectively, indicating that the prediction accuracy of MaxEnt model after parameter optimization is very high. The present suitable distribution areas of O. yunnanensis and O. delavayi predicted by this model are similar, which mainly include Sichuan, Yunnan, West Guangxi, Southeast Tibet, and West Guizhou, and their areas of suitable distribution area account for 3.52% and 4.21% of total area of China, respectively, while their areas of highly suitable distribution area account for 1.02% and 1.14% of total area of China, respectively. Under the climatic scenario of RCP8.5, potential suitable distribution area in the future of O. yunnanensis expands to the east and north, and areas of suitable distribution and highly suitable distribution areas are 8.21% and 0.30%, respectively; while potential suitable distribution area in the future of O. delavayi expands to the west and north, and areas of suitable distribution and highly suitable distribution areas are 4.41% and 1.10%, respectively. According to contribution rate, permutation importance, and Jackknife test, main climatic variables affecting distributions of O. yunnanensis and O. delavayi are variation of temperature seasonality, annual mean temperature, and annual precipitation; in addition, variance ranges of three climatic variables of O. yunnanensis in suitable distribution area are greater than those of O. delavayi when survival rate of 0.5, indicating that the ecological adaptive range of O. yunnanensis is wider. Moreover, annual mean temperature constrains the east border of distribution area of O. yunnanensis, while annual precipitation constrains its movement toward north; annual precipitation does the south border of distribution area of O. delavayi. It is suggested that the prediction result of MaxEnt model after parameter optimization is very accurate, and the predicted present suitable distribution areas of O. yunnanensis and O. delavayi are Sichuan, Yunnan, West Guangxi, Southeast Tibet, and West Guizhou, but there is difference in their potential suitable distribution areas in the future; temperature and precipitation are the main climatic factors affecting their distributions, but there are differences in main climatic factors among different distribution areas.
引文
[1] 方精云, 朱江玲, 石岳. 生态系统对全球变暖的响应[J]. 科学通报, 2018, 63(2): 136-140.
[2] 王陆军, 赵天田, 马庆华, 等. 中国特有种川榛的地理分布格局与气候环境因子的关系分析[J]. 植物资源与环境学报, 2017, 26(1): 77-83.
[3] 贾翔, 马芳芳, 周旺明, 等. 气候变化对阔叶红松林潜在地理分布区的影响[J]. 生态学报, 2017, 37(2): 464-473.
[4] 朱弘, 尤禄祥, 李涌福, 等. 浙闽樱桃地理分布模拟及气候限制因子分析[J]. 热带亚热带植物学报, 2017, 25(4): 315-322.
[5] 彭丹晓, 鲁丽敏, 陈之端. 区域生命之树及其在植物区系研究中的应用[J]. 生物多样性, 2017, 25(2): 156-162.
[6] 葛颂. 什么决定了物种的多样性? [J]. 科学通报, 2017, 62(19): 2033-2041.
[7] BELLARD C, BERTELSMEIER C, LEADLEY P, et al. Impacts of climate change on the future of biodiversity[J]. Ecology Letters, 2012, 15(4): 365-377.
[8] 刘杰, 罗亚皇, 李德铢, 等. 青藏高原及毗邻区植物多样性演化与维持机制: 进展及展望[J]. 生物多样性, 2017, 25(2): 163-174.
[9] 霍常富, 鲁旭阳, 程根伟. 中国西南山地森林演替模型的参数估计与模拟检验[J]. 东北林业大学学报, 2012, 40(10): 78-83.
[10] 张大才, 孙航. 横断山区树线以上区域种子植物的标本分布与物种丰富度[J]. 生物多样性, 2008, 16(4): 381-388.
[11] 郭兵, 姜琳, 戈大专, 等. 全球气候变暖胁迫下的雅鲁藏布江流域植被覆盖度变化驱动机制探讨[J]. 热带亚热带植物学报, 2017, 25(3): 209-217.
[12] ELITH J, KEARNEY M, PHILLIPS S. The art of modeling range-shifting species[J]. Methods in Ecology and Evolution, 2010, 1(4): 330-342.
[13] PHILLIPS S J, ANDERSON R P, SCHAPIRE R E. Maximum entropy modeling of species geographic distributions[J]. Ecological Modelling, 2006, 190(3): 231-259.
[14] PHILLIPS S J, DUDíK M. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation[J]. Ecography, 2010, 31(2): 161-175.
[15] 朱耿平, 刘强, 高玉葆. 提高生态位模型转移能力来模拟入侵物种的潜在分布[J]. 生物多样性, 2014, 22(2): 223-230.
[16] ELITH J, PHILLIPS S J, HASTIE T, et al. A statistical explanation of MaxEnt for ecologists[J]. Diversity and Distributions, 2015, 17(1): 43-57.
[17] 武晓宇, 董世魁, 刘世梁, 等. 基于MaxEnt模型的三江源区草地濒危保护植物热点区识别[J]. 生物多样性, 2018, 26(2): 138-148.
[18] 李璇, 李垚, 方炎明. 基于优化的Maxent模型预测白栎在中国的潜在分布区[J]. 林业科学, 2018, 54(8): 153-164.
[19] 朱耿平, 原雪姣, 范靖宇, 等. MaxEnt模型参数设置对其所模拟物种地理分布和生态位的影响: 以茶翅蝽为例[J]. 生物安全学报, 2018, 27(2): 118-123.
[20] WARREN D L, SEIFERT S N. Ecological niche modeling in Maxent: the importance of model complexity and the performance of model selection criteria[J]. Ecological Applications, 2011, 21(2): 335-342.
[21] SRIVASTAVA M S, KUBOKAWA T. Akaike information criterion for selecting of the mean vector in high dimensional data with fewer observations[J]. Journal of the Japan Statistical Society, 2008, 38(2): 259-283.
[22] 向其柏, 刘玉莲. 中国桂花品种图志[M]. 杭州: 浙江科学技术出版社, 2008: 9.
[23] 中国科学院中国植物志编辑委员会. 中国植物志: 第六十一卷[M]. 北京: 科学出版社, 1992.
[24] EMMANUEL R, ROTH M, ICHINOSE T, et al. Role of predictors in downscaling surface temperature to river basin in India for IPCC SRES scenarios using support vector machine[J]. International Journal of Climatology, 2010, 31(2): 159-161.
[25] KUMAR S, STOHLGREN T J. Maxent modeling for predicting suitable habitat for threatened and endangered tree Canacomyrica monticola in New Caledonia[J]. Journal of Ecology and Natural Environment, 2009, 1(4): 94-98.
[26] MUSCARELLA R, GALANTE P J, SOLEY-GUARDIA M, et al. ENMeval: an R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxent ecological niche models[J]. Methods in Ecology and Evolution, 2015, 5(11): 1198-1205.
[27] MORENO-AMAT E, MATEO R G, NIETO-LUGILDE D, et al. Impact of model complexity on cross-temporal transferability in Maxent species distribution models: an assessment using paleobotanical data[J]. Ecological Modelling, 2015, 312: 308-317.
[28] 褚建民, 李毅夫, 张雷, 等. 濒危物种长柄扁桃的潜在分布与保护策略[J]. 生物多样性, 2017, 25(8): 799-806.
[29] 刘常周, 徐波, 李志敏. 横断山区种子植物特有属的植物地理学研究[J]. 云南师范大学学报(自然科学版), 2012, 32(1): 72-78.
[30] LI G, DU S, GUO K. Evaluation of limiting climatic factors and simulation of a climatically suitable habitat for Chinese sea buckthorn[J]. PLoS One, 2015, 10(7): e131659.
[31] 孙航. 北极—第三纪成分在喜马拉雅—横断山的发展及演化[J]. 云南植物研究, 2002, 24(6): 671-688.
[32] 郑景云, 葛全胜, 赵会霞. 近40年中国植物物候对气候变化的响应研究[J]. 中国农业气象, 2003, 24(1): 28-32.
[33] 宋文静, 吴绍洪, 陶泽兴, 等. 近30年中国中东部地区植物分布变化[J]. 地理研究, 2016, 35(8): 1420-1432.
[34] 沈泽昊, 杨明正, 冯建孟, 等. 中国高山植物区系地理格局与环境和空间因素的关系[J]. 生物多样性, 2017, 25(2): 182-194.
[2] 王陆军, 赵天田, 马庆华, 等. 中国特有种川榛的地理分布格局与气候环境因子的关系分析[J]. 植物资源与环境学报, 2017, 26(1): 77-83.
[3] 贾翔, 马芳芳, 周旺明, 等. 气候变化对阔叶红松林潜在地理分布区的影响[J]. 生态学报, 2017, 37(2): 464-473.
[4] 朱弘, 尤禄祥, 李涌福, 等. 浙闽樱桃地理分布模拟及气候限制因子分析[J]. 热带亚热带植物学报, 2017, 25(4): 315-322.
[5] 彭丹晓, 鲁丽敏, 陈之端. 区域生命之树及其在植物区系研究中的应用[J]. 生物多样性, 2017, 25(2): 156-162.
[6] 葛颂. 什么决定了物种的多样性? [J]. 科学通报, 2017, 62(19): 2033-2041.
[7] BELLARD C, BERTELSMEIER C, LEADLEY P, et al. Impacts of climate change on the future of biodiversity[J]. Ecology Letters, 2012, 15(4): 365-377.
[8] 刘杰, 罗亚皇, 李德铢, 等. 青藏高原及毗邻区植物多样性演化与维持机制: 进展及展望[J]. 生物多样性, 2017, 25(2): 163-174.
[9] 霍常富, 鲁旭阳, 程根伟. 中国西南山地森林演替模型的参数估计与模拟检验[J]. 东北林业大学学报, 2012, 40(10): 78-83.
[10] 张大才, 孙航. 横断山区树线以上区域种子植物的标本分布与物种丰富度[J]. 生物多样性, 2008, 16(4): 381-388.
[11] 郭兵, 姜琳, 戈大专, 等. 全球气候变暖胁迫下的雅鲁藏布江流域植被覆盖度变化驱动机制探讨[J]. 热带亚热带植物学报, 2017, 25(3): 209-217.
[12] ELITH J, KEARNEY M, PHILLIPS S. The art of modeling range-shifting species[J]. Methods in Ecology and Evolution, 2010, 1(4): 330-342.
[13] PHILLIPS S J, ANDERSON R P, SCHAPIRE R E. Maximum entropy modeling of species geographic distributions[J]. Ecological Modelling, 2006, 190(3): 231-259.
[14] PHILLIPS S J, DUDíK M. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation[J]. Ecography, 2010, 31(2): 161-175.
[15] 朱耿平, 刘强, 高玉葆. 提高生态位模型转移能力来模拟入侵物种的潜在分布[J]. 生物多样性, 2014, 22(2): 223-230.
[16] ELITH J, PHILLIPS S J, HASTIE T, et al. A statistical explanation of MaxEnt for ecologists[J]. Diversity and Distributions, 2015, 17(1): 43-57.
[17] 武晓宇, 董世魁, 刘世梁, 等. 基于MaxEnt模型的三江源区草地濒危保护植物热点区识别[J]. 生物多样性, 2018, 26(2): 138-148.
[18] 李璇, 李垚, 方炎明. 基于优化的Maxent模型预测白栎在中国的潜在分布区[J]. 林业科学, 2018, 54(8): 153-164.
[19] 朱耿平, 原雪姣, 范靖宇, 等. MaxEnt模型参数设置对其所模拟物种地理分布和生态位的影响: 以茶翅蝽为例[J]. 生物安全学报, 2018, 27(2): 118-123.
[20] WARREN D L, SEIFERT S N. Ecological niche modeling in Maxent: the importance of model complexity and the performance of model selection criteria[J]. Ecological Applications, 2011, 21(2): 335-342.
[21] SRIVASTAVA M S, KUBOKAWA T. Akaike information criterion for selecting of the mean vector in high dimensional data with fewer observations[J]. Journal of the Japan Statistical Society, 2008, 38(2): 259-283.
[22] 向其柏, 刘玉莲. 中国桂花品种图志[M]. 杭州: 浙江科学技术出版社, 2008: 9.
[23] 中国科学院中国植物志编辑委员会. 中国植物志: 第六十一卷[M]. 北京: 科学出版社, 1992.
[24] EMMANUEL R, ROTH M, ICHINOSE T, et al. Role of predictors in downscaling surface temperature to river basin in India for IPCC SRES scenarios using support vector machine[J]. International Journal of Climatology, 2010, 31(2): 159-161.
[25] KUMAR S, STOHLGREN T J. Maxent modeling for predicting suitable habitat for threatened and endangered tree Canacomyrica monticola in New Caledonia[J]. Journal of Ecology and Natural Environment, 2009, 1(4): 94-98.
[26] MUSCARELLA R, GALANTE P J, SOLEY-GUARDIA M, et al. ENMeval: an R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxent ecological niche models[J]. Methods in Ecology and Evolution, 2015, 5(11): 1198-1205.
[27] MORENO-AMAT E, MATEO R G, NIETO-LUGILDE D, et al. Impact of model complexity on cross-temporal transferability in Maxent species distribution models: an assessment using paleobotanical data[J]. Ecological Modelling, 2015, 312: 308-317.
[28] 褚建民, 李毅夫, 张雷, 等. 濒危物种长柄扁桃的潜在分布与保护策略[J]. 生物多样性, 2017, 25(8): 799-806.
[29] 刘常周, 徐波, 李志敏. 横断山区种子植物特有属的植物地理学研究[J]. 云南师范大学学报(自然科学版), 2012, 32(1): 72-78.
[30] LI G, DU S, GUO K. Evaluation of limiting climatic factors and simulation of a climatically suitable habitat for Chinese sea buckthorn[J]. PLoS One, 2015, 10(7): e131659.
[31] 孙航. 北极—第三纪成分在喜马拉雅—横断山的发展及演化[J]. 云南植物研究, 2002, 24(6): 671-688.
[32] 郑景云, 葛全胜, 赵会霞. 近40年中国植物物候对气候变化的响应研究[J]. 中国农业气象, 2003, 24(1): 28-32.
[33] 宋文静, 吴绍洪, 陶泽兴, 等. 近30年中国中东部地区植物分布变化[J]. 地理研究, 2016, 35(8): 1420-1432.
[34] 沈泽昊, 杨明正, 冯建孟, 等. 中国高山植物区系地理格局与环境和空间因素的关系[J]. 生物多样性, 2017, 25(2): 182-194.