【作者】 张福丽；

【导师】 王志英；

【作者基本信息】 东北林业大学 ， 森林保护学， 2009， 博士

【摘要】 森林病虫害作为一类频发性生物灾害,是林业生产和生态工程建设的一个重要制约因素。我国是世界上森林病虫害发生较为严重的国家之一,经常造成危害的森林病虫有200多种。作为我国林业大省的黑龙江省,同时也是森林病虫害的重灾区。为减少森林病虫害的危害造成的损失,通过研究森林生态系统中病虫害种群变化的规律,对病虫害未来发生和增长趋势作出科学的预测预报,从而实现对森林病虫害的可持续控制。森林病虫害的发生流行是林木、有害生物和气象因素等综合作用的结果,其中气象因素是决定有害生物发生流行的关键因子。在森林生态系统中,可以根据气象因素对病虫害种群动态的影响,通过数学、生态学方法构建系统模型,利用这些模型进行主要病虫害的种群趋势的准确预测,并采取科学的综合治理措施。本论文根据数理统计原理,分析气象因子与病虫害种群变动之间的内在联系,构建了相关地区的预测模型,利用气象资料和病虫害的调查数据预报未来病虫害危害程度和发生趋势。影响森林病虫害种群动态关键气象因子的确定和预测预报模型的建立,可为黑龙江省各地、县森林病虫害的监测和防治提供的科学依据,及时掌握主要森林病虫害发生发展动态和适时为森林病虫害的防治提供科学的决策支持,最大限度地降低森林病虫害的损失,建立科学、有效、综合效益最佳的森林病虫害预测预报体系。研究了黑龙江省3种重要的森林病虫害(落叶松毛虫,杨干象,杨树溃疡病)与气象因子的关系,分别选用每年1～12月份的平均温度、平均湿度、极端高温、极端低温和降水量5个气象因子作为影响因子,通过相关分析、偏相关分析、逐步回归、多元线性回归和通径分析,得出如下结论:1、以牡丹江地区海林市、鸡西地区密山县、绥化地区庆安县和齐齐哈尔地区龙江县的气象数据和落叶松毛虫实际调查资料为依据,首次分析了落叶松毛虫预报模型。海林市逐步回归测报模型为:y=-35.374+0.286x2+0.517x72-0.694x78+0.163x80-0.154x90+1.011x73,其预报精度平均达79.73921%;密山县逐步回归测报模型为:y=-45.427+1.003x2+0.293x72+0.697x73+0.835x65+0.355x66+1.702x69,预报精度平均为88.855%,83.3%的年份精度在90%以上;庆安县逐步回归测报模型为:y=-65.303+0.925x2+0.384x65+0.316x69+0.237x72+0.047x80+0.087x59-0.008x95,预报平均精度达88.193%,精度在80%以上的年份占77.8%;龙江县逐步回归测报模型为:v=9.826+0.065x2+0.114x69+0.485x73-0.064x90-0.010x95+0.352x54,平均预报精度为78.071%。通径分析进一步揭示了模型中主要气象因子对落叶松毛虫发生的影响程度和作用方式,分析结果与实际相符。研究结果表明:海林市的上年1月平均湿度(x2)、3月平均湿度(x72)、4月极端高温(x78)、4月降水量(x80)、6月降水量(x90)、3月极端高温(x73)是影响落叶松毛虫虫口密度的主要因素;密山上年1月平均湿度(x2)、3月平均湿度(x72)、3月极端高温(x73)、1月降水量(x65)、2月平均温度(x66)、2月极端低温(x69)是主要的影响因子;庆安影响落叶松毛虫发生发展的主要因子分别为:上年1月平均湿度(x2)3月平均湿度(x72)、3月极端高温(x73)、1月降水量(x65)、2月平均温度(x66)和2月极端低温(x69);而龙江县上年1月平均湿度(x2)、2月极端低温(x69)、3月极端高温(x73)、6月降水量(x90)、7月降水量(x95)和上年11月极端低温(x54)则为主要影响落叶松毛种群消长的作用因子。综合4个县市的情况,其中上年1月平均湿度(x2)、3月平均湿度(x72)、3月极端高温(x73)和2月极端低温(x69)又处在较重要的地位。2、利用齐齐哈尔地区的拜泉县、依安县、甘南县和大庆地区的肇源县杨干象调查数据,结合各县上年10月至当年9月气象因素,通过多元线性回归建立了预测模型。研究结果表明,拜泉县影响杨干象虫口密度的主要气象因素分别为:4月极端低温(x34)、8月降水量(x55)、4月降水量(x35)、3月降水量(x30)和上年11月极端低温(x9);甘南县影响杨干象虫口密度的主要气象因素分别为:1月平均温度(x16),3月平均温度(x26),1月降水量(x20),2月极端高温(x23)和4月极端低温(x34);在肇源县影响杨干象虫口密度的主要气象因素分别为:2月极端高温(x23)、4月极端低温(x34)、6月平均温度(x41)、6月降水量(x45)、7月平均温度(x46)、7月极端低温(x49)和8月平均湿度(x52);依安县影响杨干象虫口密度的主要因素分别为:上年12月平均湿度(x12)、1月平均温度(x16)、3月平均温度(x26)、4月极端高温(x33)、6月降水量(x45)和7月极端低温(x49)。多元线性回归模型对拜泉县、甘南县、肇源县和依安县预测的平均精度分别为:95.778%、83.7%、82.9%和97.2%,其中拜泉县100%的年份、甘南县70%的年份和肇源县62.5%的年份预测精度都达到了80%以上,而依安县的预测精度最高,100%的年份预测精度都在90%以上,达到了模型预测的基本要求。综合4县市分析情况,影响杨干象发生的主要气象因子可归结为:1月平均温度(x16)、2月极端高温(x23)、3月平均温度(x26)、4月极端低温(x34)、6月降水量(x45)和7月极端低温(x49)。除6月降水量为负相关外,其余都为正相关。对这些因子的分析与杨干象的实际发生一致。构建的模型预测精度较高,达到了预测的基本要求。3、研究了绥化北林区、肇东及大庆肇源的杨树溃疡病的发生与气象因素的关系。利用逐步回归预测法建立了杨树溃疡病测报模型。结果显示,逐步回归模型对3县市的预测精度较高,北林区87.5%的年份预测精度都在90%以上、肇东和肇源的预测精度全部达到90%以上。北林区杨树溃疡病的发生与2月降水量(x5)、5月极端低温(x10)、6月平均湿度(x13)、8月降水量(x23)有关;肇东杨溃疡病的发生与2月降水量(x5)、5月极端低温(x10)、6月平均湿度(x13)、8月降水量(x23)、7月极端低温(x19)、6月平均温度(x12)关系密切;肇源影响杨树溃疡病发生的气象因子有5月极端低温(x10)、6月平均湿度(x13)、8月降水量(x23)。综合3县杨树溃疡病测报模型研究结果,5月极端低温(x10)、6月平均湿度(x13)、8月降水量(x23)是影响该病害的主要气象因子,在进行黑龙江省其它地区杨树溃疡病的预测时应着重分析这三个气象因素对杨树溃疡病的影响效应。更多还原

【Abstract】 Forest pests and diseases being as a kind of biological disaster occurring frequently are an important constrainting factor for the production of forestry and the ecological engineering construction.China is one of the countries in the world in which forest pests and diseases are more serious,and has more than 200 kinds of forest pests and diseases which often bring damage. Heilongjiang has obvious advantage at forest resource in China.At the same time,forest pests and diseases are very serious for Heilongjiang province.In order to reduce the loss because of the damage of forest pests and diseases,the future occurrence and development trends of forest pests and diseases can be forecasted,then can be controlled persistently by researching the regulation of diseases and pests populations changing in forest ecosystem.The occurrences and prevailing of forest pests and diseases are the synthesis results of trees,harmful organisms and meteorological factors,and meteorological factors are the key factors.The system model can be built by mathematics and ecology methods,and then these models can be used in forecasting population trend accurately in pest populations for important forest diseases and pests,and then to take scientific measures for the comprehensive management according to meteorological factors’ impact on the population dynamics of forest diseases and pests in forest ecosystem.The purpose of setting up mathematical models was to find out the best management measure for ecosystem by mathematical and ecological methods.In the paper,internal relations of the pests and diseases and the meteorological factors were analysed and forecasting models of relative areas were set up according to the principle of mathematical statistics,then damage degree and occurrence trend of forest pests and diseases in the future were forecasted using meteorological datum and investigation data.The scientific basis of monitoring and control to forest pests and diseases for area and county in Heilongjiang province was provided by determining key meteorological factors impacting the population dynamics of forest pests and diseases,and setting up forecasting models.The occurrence and development dynamics of major forest pests and diseases was known timely and the scientific support for the decision-making was provided timely for the prevention and control of forest pests and diseases.And to minimize the loss of forest pests and diseases,and establish a scientific and effective predicting system for forest pests and diseases with the best comprehensive benefit.The relationship of meteorological factors and three important forest pests and diseases in Heilongjiang were studied separately,which were Dendrolimus superans,Cryptorrhynchus lapathi and poplar canker.In the study,monthly average temperature,monthly average humidity,the monthly highest temperature,the monthly lowest temperature and precipitation were used as impact factors.Methods used in the paper were correlation analysis,partial correlation analysis,stepwise regression,multiple linear regression and path analysis.And the main conclusions were as follows: 1、The forecasting models of Dendrolimus superans were analyzed for the first time according to the meteorological data and the actual survey data of Dendrolimus superans which came from Hailin city of Mudanjiang area,Mishan county of Jixi area,Qing’an county of Suihua area and Longjiang county of Qiqihar region.For Hailin city,Mishan county,Qing’an county and Longjiang county,the forecasting modles were y=-35.374+0.286x2+0.517x72-0.694x78+0.163x80-0.154x90+l.011x73,y=-45.427+1.003x2+0.293x72+0.697x73+0.835x65+0.355x66+1.702x69,y=-65.303+0.925x2+0.384x65+0.316x69+0.237x72+0.047x80+0.087x59-0.008x95,and y=9.826+ 0.065x2+0.114x69+ 0.485x73-0.064x90-0.010x95+0.352x54 separately.The average forecasting accuracy of stepwise regression model was 79.73921%for Hailin city,88.855%for Mishan county, 88.193%for Qing’an county and 78.071%for Longjiang county.For Mishan county and Qing’an county,the forecasting accuracy being more than 80%was 88.3%,77.8%and 100%of the years studied respectively.Path analysis further confirmed the role of the main meteorological factors in the models affecting Dendrolimus superans.The analysises were in line with the facts.It was shown that,in Hail.in city,the average humidity of January in the last year(x2),the average humidity of March (x72),the highest temperature of April(x78),the precipitation of April(x80),the precipitation of June(x90),the highest temperature of March(x73) were the major factors affecting population density ofDendrolimus superans.And in Mishan county,they were the average humidity of January in the last year(x2),the average humidity of March(x72),the highest temperature of March(x73), the precipitation of January(x65),the average temperature of February(x66) and the lowest temperature(x69).At Qing’an county,the average humidity of January in the last year(x2),the average humidity of March(x72),the highest temperature of March(x73),the precipitation(x65) of January,the average temperature of February(x66) and the lowest temperature(x69) were the main factors affecting the occurrence and development of Dendrolimus superans,while they were the average humidity of January in the last year(x2),the lowest temperature of January(x69),the highest temperature of March(x73),the precipitation of June(x90),the precipitation of July(x95) and lowest temperature of November in the last year(x54) at Longjiang county.In conclusion,for four counties or city selected,the average humidity of January in the last year(x2),the average humidity of March(x72),the highest temperature of March(x73) and the lowest temperature of January(x69) were more important.2、When analyzing Cryptorrhynchus lapathi,Baiquan county,Yian county,Gannan county and Zhaoyuan county were used as the scopes studied,the meteorological factors from the last October to this September corresponding to the pest population density of this year and investigation data of Cryptorrhynchus lapathi was chosen.Models were established by multiple linear regression. It was shown that in Baiquan county,the main meteorological factors impacting the population density of Cryptorrhynchus lapathi were as follows:the lowest temperature of April(x34),the precipitation of August(x55),the precipitation of April(x35),the precipitation of March(x30) and the lowest temperature of the last November(x9),while the average temperature of January(x16), the average temperature of March(x26),the precipitation of January(x20),the highest temperature of February(x23) and the lowest temperature of April(x34) for Gannan.In Zhaoyuan,the main meteorological factors impacting the population density were the highest temperature of February (x23),the lowest temperature of April(x34),the average temperature of June(x41),the precipitation of June(x45),the average temperature of July(x46),the lowest temperature of July(x49) and the average humidity of August(x52),while they were the average humidity of December in the last year(x12),the average temperature of January(x16),the average temperature of March(x26),the highest temperature of April(x33)、the precipitation of June(x45),the lowest temperature of July (x49) in Yian.The average forecasting accuracy of multiple linear regression model for Baiquan county,Gannan county,Zhaoyuan county and Yian county was as follows:95.778%,83.7%,82.9% and 97.2%.And the forecasting accuracy more than 80%was 90%,70%,62.5%and 100%of the years studied respectively.All these shew that the basic requirements of model prediction has been reached.In conclusion,the main meteorological factors impacting on the occurrence and development of Cryptorrhynchus lapathi were as follows:the average temperature of January(x16),the highest temperature February(x23),the average temperature of March(x26),the lowest temperature of April(x34),the precipitation of June(x45) and the lowest temperature of July(x49).For them,the correlation coefficient were plus in addition to the precipitation of June.Analysises of these factors were consistent with the occurrence rule of Cryptorrhynchus lapathi.The forecasting accuracv of models set up was much higher,and the basic requirement was satisfied for models to forecast.3、The relationships of poplar canker and meteorological factors were analyzed for Beilin district and Zhaodong county of Suihua,and Zhaoyuan county of Daqing.The forecasting models were set up by stepwise regression using prediction factors selected.It was shown that model set up by stepwise regression displayed much higher accuracy for three counties or cities studied,and for Beilin district the forecasting accuracy being more than 90%was 87.5%,100%for Zhaodong county and Zhaoyuan county of the years studied.The occurrence of poplar canker at Beilin district had relation with the precipitation of February (x5),the lowest temperature of May(x10),the average humidity of June(x13) and the precipitation of August(x23),while the precipitation of February(x5),the lowest temperature of May(x10),the average humidity of June(x13),the precipitation of August(x23),the lowest temperature of July (x19) and the average temperature of June(x12) at Zhaodong.At Zhaoyuan the meteorological factors impacting poplar canker occurrence and development were the lowest temperature of Mav (x10),the average humidity of June(x13) and the precipitation of August(x23).In conclusion,the lowest temperature of May(x10),the average humidity of June(x13) and the precipitation of August(x23) were relatively more important for poplar canker.For other regions of Heilongjiang province,poplar canker forecasting can be mainly focused on these meteorological factors’ functions.更多还原