Reconstructing historical forest fire risk in the non-satellite era using the improved forest fire danger index and long short-term memory deep learning-a case study in Sichuan Province,southwestern China

Historical forest fire risk databases are vital for evaluating the effectiveness of past forest management approaches,enhancing forest fire warnings and emergency response capabilities,and accurately budgeting potential carbon emissions resulting from fires.However,due to the unavailability of spatial information technology,such databases are extremely difficult to build reliably and completely in the non-satellite era.This study presented an improved forest fire risk reconstruction framework that integrates a deep learning-based time series prediction model and spatial interpolation to address the challenge in Sichuan Province,southwestern China.First,the forest fire danger index(FFDI)was improved by supplementing slope and aspect information.We compared the performances of three time series models,namely,the autoregressive integrated moving average(ARIMA),Prophet and long short-term memory(LSTM)in predicting the modified forest fire danger index(MFFDI).The bestperforming model was used to retrace the MFFDI of individual stations from 1941 to 1970.Following this,the Anusplin spatial interpolation method was used to map the distributions of the MFFDI at five-year intervals,which were then subjected to weighted overlay with the distance-to-river layer to generate forest fire risk maps for reconstructing the forest fire danger database.The results revealed LSTM as the most accurate in fitting and predicting the historical MFFDI,with a fitting determination coefficient(R^2)of 0.709,mean square error(MSE)of0.047,and validation R^2 and MSE of 0.508 and 0.11,respectively.Independent validation of the predicted forest fire risk maps indicated that 5 out of 7 historical forest fire events were located in forest fire-prone areas,which is higher than the results determined from the original FFDI(2 out of 7).This proves the effectiveness of the improved MFFDI and indicates a high level of reliability of the historical forest fire risk reconstruction method proposed in this study.

Survey of damaging pests and preliminary assessment of forest health risks to the long term success of Pinus radiata introduction in Sichuan, southwest China

Pinus radiata was introduced to the summer rainfall environments of Sichuan Province, China in the 1990s as a part of an afforestation pmgram for soil and water conservation in the arid and semi-arid river valley area of Aba Prefecture. Within this region a total area of 26 000 ha have been identified through climate matching as suitable and a further 63 000 ha potentially suitable for environmental plantings of P. radiata. The plantations are being established in widely separated small patches on steep and degraded slopes along the dry river valley The newly introduced P. radiata are exposed to two kinds of forest health risks： they may be attacked by （a） indigenous pathogens and pests against which they may not possess any resistance or （b） by inadvertently introduced foreign pests or pathogens. This paper presents a survey of the potential damaging pests and a preliminary assessment of forest health risks facing the P. radiata plantations over a much longer timeframe than the initial phase of introduction and early plantation establishment. An empirical appmach was adopted to evaluate forest health risks by a combination of literature review, examination of historical records of pest and disease outbreaks in the surrounding conifemus forests, field surveys and inspections, specimen collection and identification, and most importantly, expert analysis of the likelihood of attack by specific pests and pathogens and the subsequent impact of such attacks. The assessment identified some specific forest health risks to the long-term success of P. radiata introduction in this area. These risks are closely associated with the indigenous pests and pathogens of the two native pine species, P. tabulaeformis and P. armondii since these pests and pathogens are considered more likely to establish on P. radiata over time. Exotic pests and pathogens are of a quarantine concern at present. Based on the results of assessment, recommendations are pmvided to improve forest vigour and to reduce the forest health risks posed by indigenous as well as exotic pests and pathogens to the introduced P. radiata. Ways to increase the ability to manage the forest health risks once a particular pest infestation and disease eventuates are also recommended. Although detrimental to the survival and growth of the introduced P. radiata, the impact of identified forest health risks are not considered to be fatal to the long term success ofP. radiata in this area.

Forest fire risk zone mapping from satellite images and GIS for Baihe Forestry Bureau, Jilin, China

A forest fire can be a real ecological disaster regardless of whether it is caused by natural forces or human activities, it is possible to map forest fire risk zones to minimize the frequency of fires, avert damage, etc. A method integrating remote sensing and GIS was developed and applied to forest fire risk zone mapping for Baihe forestry bureau in this paper. Satellite images were interpreted and classified to generate vegetation type layer and land use layers （roads, settlements and farmlands）. Topographic layers （slope, aspect and altitude） were derived from DEM. The thematic and topographic information was analyzed by using ARC/INFO GIS software. Forest fire risk zones were delineated by assigning subjective weights to the classes of all the layers （vegetation type, slope, aspect, altitude and distance from r3ads, farmlands and settlements） according to their sensitivity to fire or their fire-inducing capability. Five categories of forest fire risk ranging from very high to very low were derived automatically. The mapping result of the study area was found to be in strong agreement with actual fire-affected sites.

RS AND GIS-BASED FOREST FIRE RISK ZONE MAPPING IN DA HINGGAN MOUNTAINS

The Da Hinggan Mountains is one of the most important forest areas in China, but forest fire there is also of high frequency. So it is completely necessary to map forest fire risk zones in order to effectively manage and protect the forest resources. Two forest farms of Tuqiang Forest Bureau (53 degrees 34'-52 degrees 15'N,124 degrees 05'-122 degrees 18'E) were chosen as typical areas in this study. Remote sensing (RS) and Geographic Information System (GIS) play a vital role and can be used effectively to obtain and combine different forest-fire-causing factors for demarcating the forest fire risk zone map. Forest fire risk zones were described by assigning subjective weights to the classes of all the coverage layers according to their sensitivity to fire, using the ARC/INFO GIS software. Four classes of forest fire risk ranging from low to extremely high were generated automatically in ARC/INFO. The results showed that about 60.33% of the study area were predicted to be upper moderate risk zones, indicating that the forest fire management task in this area is super onerous. The RS and GIS-based forest fire risk model of the study area was found to be highly compatible with the actual fire-affected sites in 1987. Therefore the forest fire risk zone map can be used for guidance of forest fire management, and as basis for fire prevention strategies.

Comparative analysis of multi-criteria probabilistic FR and AHP models for forest fire risk(FFR)mapping in Melghat Tiger Reserve(MTR)forest

A comparative study of Frequency Ratio(FR)and Analytic Hierarchy Process(AHP)models are performed for forest fire risk(FFR)mapping in Melghat Tiger Reserve forest,central India.Identification of FFR depends on various hydrometeorological parameters altitude,slope,aspect,topographic position index,normalized differential vegetation index,rainfall,air temperature,land surface temperature,wind speed,distance to settlements,and distance by road are integrated using a GIS platform.The results from FR and AHP show similar trends.The FR model was significantly higher accurate(overall accuracy of 81.3%,kappa statistic 0.78)than the AHP model(overall accuracy 79.3%,kappa statistic 0.75).The FR model total forest fire risk areas were classified into five classes:very low(7.1%),low(22.2%),moderate(32.3%),high(26.9%),and very high(11.5%).The AHP fire risk classes were very low(6.7%),low(21.7%),moderate(34.0%),high(26.7%),and very high(10.9%).Sensitivity analyses were performed for AHP and FR models.The results of the two different models are compared and justified concerning the forest fire sample points(Forest Survey of India)and burn images(2010-2016).These results help in designing more effective fire management plans to improve the allocation of resources across a landscape framework.

An easy,accurate and efficient procedure to create forest fire risk maps using the SEXTANTE plugin Modeler

To prevent, detect, and protect against forest fires, forest personnel need to define rules for determining forest fire risk. In Portugal, all municipalities must annually produce forest fire risk （FFR） maps. To produce more reliable FFR maps more easily, we developed an open source model using the Modeler plugin of SEXTANTE in the program QGIS version 2.0 Dufour. The model provides all the maps involved in the FFR model （susceptibility map, hazard map, vulnerability map, economic value map, and potential loss map） and was produced according to Portuguese Forest Authority＇s （AFN, Autoridade Florestal Nacional） rules for determining the FFR. This model was tested for the Portuguese municipality Santa Maria da Feira, where 40 % of the total municipality area falls in the category ＂very high＂ or ＂high＂ fire risk. The ＂very high＂ fire risk area is mainly classified as broad-leaved forest and has the steepest slopes （〉15 %）. The distance of burned areas to roads was also analyzed; the proportion of burned areas increased with increasing distance to the main roads. In addition, 92.6 % of the ＂high＂ and ＂very high＂ risk zones were located in areas with lower elevation. These results confirmed that forest fire is strongly influenced not only by environmental factors but also by anthropogenic factors. The procedure implemented here was compared with our open source application already available in QGIS and also to the same procedure implemented in GIS pro- prietary software. Although the results were obviously the same, the model developed here presents several advan- tages over the other two approaches. Besides being faster, it is easy to change the model parameters according to user needs （i.e., to the rules of different countries）, and can be modified and adapted to other variables and other areas to create risk maps for different natural phenomena （e.g., floods, earthquakes, landslides）. The model is easy to use and to create risk and hazard maps rapidly in a free, open source environment that does not require any programming knowledge.

Forest fire risk assessment in parts of Northeast India using geospatial tools

Forest fire is a major cause of changes in forest structure and function. Among various floristic regions, the northeast region of India suffers maximum from the fires due to age-old practice of shifting cultivation and spread of fires from jhum fields. For proper mitigation and management, an early warning of forest fires through risk modeling is required. The study results demonstrate the potential use of remote sensing and Geographic Information System （GIS） in identifying forest fire prone areas in Manipur, southeastern part of Northeast India. Land use land cover （LULC）, vegetation type, Digital elevation model （DEM）, slope, aspect and proximity to roads and settlements, factors that influence the behavior of fire, were used to model the forest fire risk zones. Each class of the layers was given weight according to their fire inducing capability and their sensitivity to fire. Weighted sum modeling and ISODATA clustering was used to classify the fire zones. TO validate the results, Along Track Scanning Radiometer （ATSR）, the historical fire hotspots data was used to check the occurrence points and modeled forest fire locations. The forest risk zone map has 55-63% of agreement with ATSR dataset.

GIS, Forest Fire Prevention and Risk Matrix in the National Forest of Khoudida, Sidi Bel Abbes, Algeria

Forest fires in Algeria are ravaging an average of more than 32,000 hectares annually despite the prevention and control plan put in place. They are the most damaging factor of degradation of the forest and weigh heavily on the environment and the local economy. Conventional methods for fire prevention and control are time consuming and are not always reliable in view of the complexity and diversity of forest ecosystems. The main idea behind this study is to use the GIS and remote sensing for the development of a fire risk map of the Khoudida State Forest (Algeria). The approach adopted involves three parameters that control the fire behavior, which are: the top-morphology of the field, the combustibility of the plant cover and hazards. For each factor its correlation with risk was evaluated;the combination of the slope, altitude and exposure parameters in the topo-morphological index and the hazard map made it possible to evaluate the average risk for an area of more than 2132 hectares, 1521 hectares high and only 493 hectares, respectively 51.4%, 36.7% and 11.9%.

Forest fire risk indices and zoning of hazardous areas in Sorocaba,Sao Paulo state,Brazil

This study compares the performance of three fire risk indices for accuracy in predicting fires in semideciduous forest fragments,creates a fire risk map by integrating historical fire occurrences in a probabilistic density surface using the Kernel density estimator(KDE)in the municipality of Sorocaba,Sao Paulo state,Brazil.The logarithmic Telicyn index,Monte Alegre formula(MAF)and enhanced Monte Alegre formula(MAF+)were employed using data for the period 1 January 2005 to 31 December 2016.Meteorological data and numbers of fire occurrences were obtained from the National Institute of Meteorology(INMET)and the Institute for Space Research(INPE),respectively.Two performance measures were calculated:Heidke skill score(SS)and success rate(SR).The MAF+index was the most accurate,with values of SS and SR of 0.611%and 62.8%,respectively.The fire risk map revealed two most susceptible areas with high(63 km^2)and very high(47 km^2)risk of fires in the municipality.Identification of the best risk index and the generation of fire risk maps can contribute to better planning and cost reduction in preventing and fighting forest fires.

Study on Forecasting Method of Forest Fire Risk Grade in Putian City, China

From January 1, 2014, the basic stations of meteorological observation countries have changed from small evaporation observations to large-scale evaporation observations. National general weather stations have canceled observations on evaporation, but small evaporation is very important for forest fire risk prediction. In order to make the prediction of forest fire risk level objectively, weather data in Putian City, China and the multi-linear regression analysis method is used to calculate the daily evaporation amount in the more advanced SPSS16.0 software (English version), and the data of the last 5 years of each site are selected and fitted. Results showed that we accurately calculated the evaporation of the next day to make up for the lack of data due to the adjustment of the evaporation observation project. According to the forest fire risk weather index corresponding to many meteorological factors such as evaporation, temperature, humidity, sunshine and wind speed, the forest fire risk meteorological grade standard was designed to make a more accurate forest fire risk grade forecast.

融合贝叶斯优化随机森林的机场旅客风险评估研究

为科学有效地评价旅客风险,提高旅客出行便捷性,以机场离港旅客为研究对象,借助旅客姓名记录获取评价旅客风险的相关信息,结合旅客安检信息,编制问卷进行调查。运用SPSS 22.0软件对有效问卷数据进行合理性检验,构建民航旅客风险评价指标体系;在此基础上,利用贝叶斯优化随机森林模型对旅客风险等级进行综合评价。结果表明:在影响旅客风险等级划分的8个指标中,年飞行次数和托运行李违禁品记录重要性最大,性别重要性最小;相较于多种传统的算法,贝叶斯优化随机森林的分类性能更高,准确率达到97%。研究结果对机场实施旅客分类安检具有一定的指导作用。

基于随机森林算法和逻辑回归预测妊娠剧吐的危险因素

目的开发新的预测模型,识别妊娠剧吐危险因素,以有效预测妊娠剧吐发病。方法选取2020至2022年118例孕妇作为研究对象,分为妊娠剧吐组和非妊娠剧吐组,每组59例。构建一个随机森林模型和传统的多因素逻辑回归模型。然后分别分析2个模型的分析结果。结果随机森林分析结果显示在以基尼系数平均下降为基准,影响因素排序为:低BMI(BMI<20 kg/m^(2)),妊娠剧吐家族史,孕妇年龄,胎儿性别,葡萄胎病史等。多因素逻辑回归模型对数据集的分析结果确定了包括低BMI、妊娠剧吐家族历史、女性胎儿、吸烟史、多胎妊娠、精神病史、糖尿病、上消化道病史、甲状腺机能亢进史和葡萄胎病史均与妊娠剧吐发生显著相关。结论本研究建立了基于随机森林的发生风险预测模型,具有较好的预测价值,为预防诊断妊娠剧吐提供科学依据和快速筛查工具。

基于随机森林算法建立非急诊大手术后延迟拔管的预测模型

目的基于随机森林算法分析非急诊大手术后延迟拔管的影响因素,建立并验证术后延迟拔管的预测模型。方法回顾性收集2018年1月至2022年12月全麻下行非急诊大手术的7528例患者的临床资料。根据术后2 h内是否拔管,将患者分为两组:非延迟拔管组(≤2 h)和延迟拔管组(>2 h)。将患者按照7∶3分为训练集和验证集,通过LASSO回归、Logistic回归筛选术后延迟拔管的预测因素,采用随机森林算法建立并验证预测模型。结果有123例(1.6%)出现术后延迟拔管。ASA分级、科室、术中使用氟比洛芬酯、右美托咪定、激素、术中出现低钙血症、重度贫血、术中输血、气道痉挛是术后延迟拔管的独立预测因素。基于随机森林算法建立的预测模型在验证集中的曲线下面积(AUC)为0.751(95%CI 0.742~0.778),敏感性98.1%,特异性41.9%。结论基于随机森林算法建立的非急诊大手术后拔管延迟的预测模型具有较好的预测性能,利用该模型有助于预防非急诊大手术后延迟拔管。

河南省森林生态系统牛膝菊入侵风险评价

牛膝菊(Galinsoga parviflora Cav.)适应能力强,扩展速度快,量大,难以根除,对森林生态系统破坏大。通过历史资料查阅及野外调查,并结合其生物学、生态学特性,对河南省森林生态系统牛膝菊入侵风险定量分析,计算出风险值为2.09,属于中风险入侵植物,会造成入侵地生物多样性降低,破坏当地森林生态系统,应严加防范。

中国农田磷流失风险评价及其关键驱动因素

农田面源磷流失是农业面源污染的重要原因之一,识别流域内农田磷流失风险的关键源区及其影响因子是面源污染控制的重要手段。基于磷指数模型开展2000—2020年中国农田磷流失风险评估,选取土壤有效磷含量、磷肥施用量为源因子,土壤侵蚀模数、年径流深、农田和水体间归一化距离指数为迁移因子,结合GIS技术评估了中国农田磷流失的关键源区;在此基础上,利用随机森林法分析影响中国农田磷流失的关键因子,并通过结构方程模型揭示了农田磷流失风险指数与各因子的关系。结果表明:1)2000—2020年中国农田的磷流失的低、中、高、极高风险面积分别占农田总面积的43.8%、40.5%、13.4%、2.4%。2)中国农田磷流失在2000、2005、2010、2015、2020年高风险和极高风险总面积的年平均占比从大到小依次为:淮河流域、长江流域、珠江流域、东南诸河流域、松辽河流域、西南诸河流域、黄河流域、内陆河流域、海河流域。3)影响农田磷流失风险的关键源因子和迁移因子分别为土壤有效磷含量和归一化距离指数,其重要性特征值分别为129.53和65.12,土壤有效磷含量是农田磷流失最主要影响因子。4)磷流失风险指数与源因子指数、迁移因子指数呈极显著正相关,选取的14个指标对磷指数的解释度达0.62,其中源因子和迁移因子对磷指数的贡献率分别为0.77、0.19(P <0.001)。研究结果可为中国农田磷流失风险评估提供科学参考,对中国农业面源污染的宏观防控及战略决策具有重要意义。

内蒙古森林火灾发生风险及其驱动因素

森林火灾是危害森林健康的主要灾害之一,科学预测森林火灾是预防森林火灾的重要依据。以中国新旧森林火灾政策作为分界线,将内蒙古森林历史火灾数据分为四个阶段,基于增强回归树模型建立内蒙古森林火灾发生模型,预测森林火灾发生情况,解释不同时期森林火灾和火灾风险变化的差异。预测结果表明:(1)4个时期建模精度AUC均大于0.94,表明BRT模型能够较好地预测研究区森林火灾的发生;(2)气温日较差、日最小相对湿度、上一年春防累计降水量、上一年秋防累计降水量、上一年春防最高地表气温海拔、距火点最近公路距离被确定为影响内蒙古森林火灾发生的重要驱动因素。(3)新旧《森林防火条例》实施前后森林火灾风险等级变化:1981—1988年3月14日,中、高和极高森林火灾风险区分布在呼伦贝尔的东部,而2009—2020年中、高和极高森林火灾风险区分布在呼伦贝尔南部和中部、赤峰市的西南部、锡林郭勒盟和呼和浩特市的中部、乌兰察布市和包头市的南部以及鄂尔多斯市的东部。有助于了解不同时期《森林防火条例》影响下的内蒙古森林火灾的驱动因素和火险等级的变化,为优化森林火灾管理政策及预测预报工作提供科学依据。

基于人工智能方法的隧道塌方风险预测研究

为了对隧道塌方风险展开研究,整理246起隧道塌方事故案例,通过建立塌方风险评估指标体系,基于人工智能预测方法,分别采用随机森林算法、径向基函数神经网络、BP神经网络模型、粒子群算法优化BP神经网络模型,对塌方风险进行预测。结果表明,随机森林算法、径向基函数神经网络、BP神经网络模型、粒子群算法优化BP神经网络模型的塌方预测准确率分别为81.67%、83.33%、86.67%、93.33%,F_(1)值分别为0.645、0.642、0.5、0.833。粒子群算法优化BP神经网络模型预测准确率和F_(1)值均大幅提高,预测效果最好,大大减少了评估结果的主观性,为隧道塌方风险研究提供了新的研究思路。

基于随机森林算法的重型颅脑损伤患者并发急性胃肠损伤的现状及风险模型构建

目的探讨重型颅脑损伤患者并发急性胃肠损伤的危险因素,为预防急性胃肠损伤提供借鉴。方法2021年1月至2023年1月,便利抽样法选取某院收治的重型颅脑损伤患者150例为研究对象,建立基于重型颅脑损伤并发急性胃肠损伤的危险因素的随机森林算法的预测模型。结果150例重症颅脑损伤患者中,并发急性胃肠损伤患者94例,占62.67%。是否并发急性胃肠道损伤的患者在糖尿病、白蛋白、APACHE-Ⅱ评分、休克指数、液体负平衡、酸中毒、深度镇静、呼吸衰竭方面的差异均有统计学意义(均P<0.05)。构建重型颅脑损伤并发急性胃肠损伤的随机森林模型,树的数量为103时出现的错误率最低;影响重型颅脑损伤并发急性胃肠损伤的因素重要性排序为糖尿病、液体负平衡、急性生理与慢性健康评分、白蛋白、深度镇静及酸中毒;随机森林模型预测重型颅脑损伤并发急性胃肠损伤的受试者工作特征曲线(receiver operating characteristic,ROC)下面积(area under curve,AUC)为0.798,Logistic回归模型的AUC为0.773。结论构建的重型颅脑损伤并发急性胃肠损伤的风险预测模型预测效能较高,临床值得推广应用。

森林保险实施情况、发展困境与优化策略——以河南省为例

我国自2009年起实施森林保险中央财政保费补贴试点政策,森林保险参保面积和保费规模不断扩大,河南省自2012年起开始实施森林保险试点工作。以河南省为例,总结森林保险发展情况和目前的发展困境。研究发现:自试点以来,河南省森林保险覆盖率稳步上升,但公益林保险与商品林保险发展相对不均衡,商品林保险参保面积和参保率长期处于较低水平,现行森林保险产品难以满足实际风险管理需要,森林保险整体赔付率较低,保险风险补偿功能和产品创新能力亟待提升。因此,从优化森林保险运行机制、创新保险产品、改进保险定价方式、提高保险服务质量等方面提出对策建议,为河南省及全国森林保险高质量发展提供参考。

气象要素与潜在林火风险关联性研究

为探究川西南地区复杂气候与森林火灾之间的关联性,以四川省凉山彝族自治州木里藏族自治县(以下简称“四川省凉山州木里县”)为例,通过局部规模的气象异常值变化,分析了各气象要素与森林火灾的潜在联系性。再通过FWI与FFDI,分析木里县历史火险变化,验证了其与木里县森林火灾的关联性。结果表明:森林火灾发生前后气压、气温、湿度、风速、蒸发量出现不同周期的连续增加(或衰减)。同时,木里县森林火险指数FWI与FFDI呈现出年际震荡且逐渐增大的趋势,且木里县林火频数与FWI的年均相关系数(46.9%)、月均相关系数(80.7%)均大于FFDI(年相关系数38.7%、月相关系数53.1%)。此外,林火风险值出现极端异常还与同期(或同年)发生的气候事件密切相关。因此,气候异常事件所导致的持续高温、低湿、多风的气象条件是森林火险等级快速升高或重大森林火灾事故的本质原因之一。