﻿ 黔南地区气象因子与森林火灾发生次数之间的关系
 林业科学  2011, Vol. 47 Issue (10): 128-133 PDF
0

#### 文章信息

Xiao Yundan, Ju Hongbo, Zhang Xiongqing, Ji Ping

Relationship between Fire-Danger Weather and Forest Fire in Qiannan Area

Scientia Silvae Sinicae, 2011, 47(10): 128-133.

### 作者相关文章

Relationship between Fire-Danger Weather and Forest Fire in Qiannan Area
Xiao Yundan, Ju Hongbo, Zhang Xiongqing, Ji Ping
Institute of Forest Resources Information Techniques, CAF Beijing 100091
Abstract: In this study, based on data of the forest fire occurrence and meteorological variables in spring fireproofing period in Qiannan area, Poisson regression model, negative binomial model, zero-inflated negative binomial model and Hurdle model were respectively employed to predict the forest fires under fire-danger climate, and those models were compared with each other based on the prediction. The results showed that: Poisson regression model did not fit well into the over-dispersion data. Negative binomial distribution fitted better into the data than Poisson distribution. But both of them were not suitable for simulating zero drived dispersion data. Zero-inflated negative binomial regression model and Hurdle model were useful methods for such data. Zero inflated negative binomial regression model and Hurdle model performed better than other two models in predicting forest fires. Moreover, Hurdle model was even superior to zero-inflated negative binomial model.
Key words: forest fire    fire-danger weather    Poisson regression model    negative binomial model    zero-inflated negative binomial model    Hurdle model

1 研究区概况与数据来源 1.1 研究区概况

1.2 数据来源

2 数据分析 2.1 零数据分布

 图 1 森林火灾发生数直方图 Figure 1 Histogram of forest fires
2.2 多重共线性检验

2.3 模型评价

 (1)
 (2)

3 研究方法与结果

3.1 Poisson模型的拟合结果分析

Poisson回归模型是一种常用的离散数据计算方法(恽振先，1992Bailer et al., 1997Mandallaz et al., 1997)。由于Poisson模型和负二项模型比较常见，文中未一一列举。根据模型的参数统计t检验结果，剔除掉不显著的变量，最终得到Poisson模型的参数估计值及评价统计量(表 3)。

3.2 负二项回归模型的拟合结果分析

3.3 零膨胀负二项模型及拟合结果分析

 (3)
 (4)

3.4 Hurdle模型及拟合结果分析

 (5)

 (6)

4 模型检验

 图 2 Hurdle模型的实际值与预测值 Figure 2 Predicted value of Hurdle model vs. observed value
5 结论

 郭福涛. 2010. 基于负二项和零膨胀负二项回归模型的大兴安岭地区雷击火与气象因素的关系[J]. 植物生态学报, 34(5): 571-577. 宋卫国, 马剑, SatohK, 等. 2006. 森林火险与气象因素的多元相关性及其分析[J]. 中国工程科学, 8(2): 61-66. 徐飞. 2009. 负二项回归模型在过离散型索赔次数中的应用研究[J]. 统计教育, (4): 53-55. 恽振先. 1992. Poisson回归模型估计启东肝病高危人群14年随访资料的PAR[J]. 中国公共卫生学报, 11(4): 197-200. 曾平, 刘桂芬, 曹红艳. 2008. 零膨胀模型在心肌缺血节段数影响因素研究中的应用[J]. 中国卫生统计, 25(5): 464-466. 张尚印, 祝昌汉, 陈正洪. 2000. 森林火灾气象环境要素和重大林火研究[J]. 自然灾害学报, 9(2): 111-117. Bailer A J, Reed L D, Stayner L T. 1997. Modeling fatal injury rates using Poisson regression: A case study of workers in agriculture, forestry, and fishing[J]. Journal of Safety Research, 28(3): 177-186. DOI:10.1016/S0022-4375(97)80006-0 Bradstock R A, Gill A M. 1998. Bushfire risk at the urban interface estimated from historical weather records: consequences for the use of prescribed fire in the Sydney region of south-eastern Australia[J]. Journal of Environmental Management, 52(3): 259-271. Cameron A C, Trivedi P K. 2005. Microeconometrics: Methods and Applications[M]. Cambridge: Cambridge University Press. Eskelson B N I, Temesgen H, Barrett T M. 2009. Estimating cavity tree and snag abundance using negative binomial regression models and nearest neighbor imputation methods[J]. Canadian Journal of Forest Research, 39(9): 1749-1765. DOI:10.1139/X09-086 Greene W. 2008. Functional forms for the negative binomial model for count data[J]. Economics Letters, 99(3): 585-590. DOI:10.1016/j.econlet.2007.10.015 Ground M. 2007. Hurdle models of alcohol and tobacco expenditure in South African households[J]. South African Journal of Economics, 76(1): 132-143. Jones A M. 1989. A Double-hurdle model of cigarette consumption[J]. Journal of Applied Econometrics, 4(1): 23-39. DOI:10.1002/(ISSN)1099-1255 Kagraoka Y. 2005. Modeling Insurance Surrenders by the Negative Binomial Model. (2005-06-03). http://www.musashi.jp/~kagraoka/research/NBM_017.pdf. Lambert D. 1992. Zero-inflated Poisson regression, with an application to defects in manufacturing[J]. Techno Metrics, 34(1): 1-14. DOI:10.2307/1269547 Lee B S, Alexander M E. 2002. Information systems in support of wild land fire management decision making in Canada[J]. Computers and Electronics in Agriculture, 37(3): 185-198. MacNeil M A, Carlson J K, Beerkircher L R. 2009. Shark depredation rates in pelagic longline fisheries: a case study from the Northwest Atlantic[J]. ICES Journal of Marine Science, 66(4): 708-719. DOI:10.1093/icesjms/fsp022 Mandallaz D, Ye R. 1997. Prediction of forest fires with Poisson models[J]. Canadian Journal of Forest Research, 27: 1685-1694. DOI:10.1139/x97-103 McCullagh P, Nelder J. 1989. Generalized Linear Models[M]. 2nd ed. Boca Raton, FA, USA: CRC Press: 511. Moffatt P G. 2005. Hurdle models of loan default[J]. Journal of the Operational Research Society, 56(9): 1063-1071. DOI:10.1057/palgrave.jors.2601922 Mullahy J. 1986. Specification and testing of some modified count data models[J]. Journal of Econometrics, 33(3): 341-365. DOI:10.1016/0304-4076(86)90002-3 Pausas J G. 2004. Changes in fire and climate in the eastern Iberian Peninsula (Mediterranean Basin)[J]. Climatic Change, 63(3): 337-350. DOI:10.1023/B:CLIM.0000018508.94901.9c Vuong Q H. 1989. Likelihood ratio tests for model selection and non-nested hypotheses[J]. Econonmetrica, 57(2): 307-333. DOI:10.2307/1912557 Welsh A H, Cunningham R B, Donnelly C F, et al. 1996. Modelling the abundance of rare species: statistical models for counts with extra zeros[J]. Ecological Modeling, 88(10): 297-308. Zeileis A, Kleiber C, Jackman S. 2008. Regression models for count data in R[J]. Journal of Statistical Software, 27(8): 1-25.