Remote sensing-modelisation approach for diurnal estimation of burnt biomass in the Central African Republic Savanna

Experimental studies and mesoscale modeling of atmospheric chemistry require a good knowledge of the sources of the atmospheric constituent, at a temporal scale of about one hour and at a spatial scale corresponding to the model grid. A combined remote sensing/modeling approach for the estimation of the diurnal distribution of the amount of biomass burning in Central African Republic (C.A.R.) savanna fires is proposed. The fire propagation model (BEHAVE) developed by Rothermel was adapted to the fuel characteristics encountered in C.A.R. Ground and airborne measurements with satellite images (NOAA/AVHRR) were used to predict an accurate estimate of the burnt biomass. This combination allows the calculation of the distribution of the number of fires during the day providing an evaluation of the instantaneous fluxes of the compounds emitted in the atmosphere by these fires.     

极轨气象卫星森林火灾实时监测系统

介绍了新开发的极轨气象卫星林火实时监测系统的原理、功能及其数据处理流程.其特点是在卫星过境的同时就可处理出森林火灾信息,解决了目前同类气象卫星林火监测系统不能实时监测的问题.极轨气象卫星林火实时监测系统是由3台计算机组成的并行处理系统.该系统将火点判识的时序从模糊到确定分3步进行,并将数据接收、火点判断、火点定位、显示等的运行时序分配在3台计算机中同步运行.该系统于2005年在广州气象卫星地面站投入业务运行.与原有卫星遥感林火监测系统比较,该系统对森林火灾监测的时间可提早约30 min.     

Fire activity in Borneo driven by industrial land conversion and drought during El Niño periods, 1982–2010

Tropical rainforests, naturally resistant to fire when intact, are increasingly vulnerable to burning due to ongoing forest perturbation and, possibly, climatic changes. Industrial-scale forest degradation and conversion are increasing fire occurrence, and interactions with climate anomalies such as El Niño induced droughts can magnify the extent and severity of fire activity. The influences of these factors on fire frequency in tropical forests has not been widely studied at large spatio-temporal scales at which feedbacks between fire reoccurrence and forest degradation may develop. Linkages between fire activity, industrial land use, and El Niño rainfall deficits are acute in Borneo, where the greatest tropical fire events in recorded history have apparently occurred in recent decades. Here we investigate how fire frequency in Borneo has been influenced by industrial-scale agricultural development and logging during El Niño periods by integrating long-term satellite observations between 1982 and 2010 – a period encompassing the onset, development, and consolidation of its Borneo’s industrial forestry and agricultural operations as well as the full diversity of El Niño events. We record changes in fire frequency over this period by deriving the longest and most comprehensive spatio-temporal record of fire activity across Borneo using AVHRR Global Area Coverage (GAC) satellite data. Monthly fire frequency was derived from these data and modelled at 0.04° resolution via a random-forest model, which explained 56% of the monthly variation as a function of oil palm and timber plantation extent and proximity, logging intensity and proximity, human settlement, climate, forest and peatland condition, and time, observed using Landsat and similar satellite data. Oil-palm extent increased fire frequency until covering 20% of a grid cell, signalling the significant influence of early stages of plantation establishment. Heighted fire frequency was particularly acute within 10 km of oil palm, where both expanding plantation and smallholder agriculture are believed to be contributing factors. Fire frequency increased abruptly and dramatically when rainfall fell below 200 mm month⁻¹, especially as landscape perturbation increased (indicated by vegetation index data). Logging intensity had a negligible influence on fire frequency, including on peatlands, suggesting a more complex response of logged forest to burning than appreciated. Over time, the epicentres of high-frequency fires expanded from East Kalimantan (1980’s) to Central and West Kalimantan (1990’s), coincidentally but apparently slightly preceding oil-palm expansion, and high-frequency fires then waned in East Kalimantan and occurred only in Central and West Kalimantan (2000’s). After accounting for land-cover changes and climate, our model under-estimates observed fire frequency during ca. 1990–2002 and over-estimates it thereafter, suggesting that a multi-decadal shift to industrial forest conversion and forest landscapes may have diminished the propensity for high-frequency fires in much of this globally significant tropical region since ca. 2000.     

环境减灾卫星在甘肃省草原火灾监测中的应用研究

通过分析环境减灾卫星（HJ）的光谱特征,参考MODIS火情监测模型,发展了基于HJ卫星的火情遥感监测方法。针对甘肃省2011年2月10日发生在玛曲、2012年12月8日发生在岷县境内的2次草原火灾,研究了HJ卫星相应通道的本地化阈值,采用植被指数的多时相阈值法提取过火范围并计算了过火面积,同时与甘肃草原监理站的过火面积观测资料及同期MODIS的监测结果分别进行比对。结果表明,2次判定火点的阈值为T4〉310K;HJ卫星监测到的玛曲、岷县的火灾面积分别为65．3hm2、19．8hm2,监测精度为98％和90％,较MODiS监测精度提高了10．5％和2g．9％;Hj卫星进行火情遥感监测的空间分辨率更高,监测结果更精细,具有很好的业务化应用价值。     

河北省森林火险预报模式的改进及应用

采用数理统计的方法,增加了MODIS卫星遥感监测的土壤湿度和热源点预报因子,对河北省气象局原有的森林火险预报模式进行了改进,针对不同区域分别建立了森林火险预报模式,并对2012年河北省森林火灾实际发生情况进行了分析和应用效果检验。结果表明,2012年防火期,实况出现火灾,改进的火险模式预报森林火险气象等级5级为预报完全正确比率达66．3％;预报火险气象等级为4级（高度火险）及以上的正确率达83．1％;预报火险气象等级为3级（中度火险）及以上的正确率达98．8％;在所有的预报样本中,森林火险气象等级预报5级,但实况没有出现火灾的空报率为6．8％。检验结果显示,改进后的森林火险预报模式的应用效果更接近实际情况。     

卫星微波被动遥感云中液态水的研究进展

卫星微波被动遥感云中液态水的研究在近30年中取得了较大的进展.卫星遥感仪器的不断进步、探测精度的不断提高为卫星微波被动遥感云中液态水提供了技术和资料保障,在此基础上针对卫星微波被动遥感云中液态水技术而发展的一些统计反演方法和物理反演方法不断地得到改进、完善和提高.文中主要就各个时期发展的主要反演方法做了简要回顾,概括了这些反演方法的关键技术,评述了各自的优点和局限性.由于海表和陆表的辐射特性的巨大差异,分别就海洋上空和陆地上空云中液态水的统计和物理反演技术方法进行了综述,并进行了比较.针对地表比辐射率这一制约陆地上空云中液态水反演技术发展的瓶颈问题,也进行了相关评述和讨论.     

基于NOAA数据的近年贵州野火时空分布特征分析

本文利用2007～2017年贵州省卫星遥感火险监测数据,对近11年来贵州野外火灾发生的时空分布特征静下来分析,结果如下：1.春季是四季中野火发生频次最高的季节,占了全年的近一半,其次是冬季; 2月份野火发生频次最高,其次是3月份,2～5月发生的野火次数占了全年72.6%。2.近11年遥感监测野火频次总体呈逐年减少的趋势,平均每年减少194起,期间呈现分为2个变化阶段,2007～2010年处于上升趋势,2011～2017年呈现减少趋势。3.从空间分布来看,9个市州中发生野火频次最高的是黔南,占全省的21.3%,黔西南排第二,占全省的19.7%,其后依次是毕节、安顺、黔东南、六盘水、遵义、铜仁、贵阳;森林火灾占总野火频次的一半。     

一种由卫星遥感资料获得的修正的土壤调整植被指数

目前,通过卫星遥感资料确定区域面上植被分布、类型的研究受到许多实用领域的普遍重视,并由此提出了许多形式不同的植被指数。由于土壤背景噪声是造成植被指数不确定的重要原因之一,为此不同学者在标准化差值植被指数的基础上提出了多种旨在能消弱土壤背景噪声的土壤调整植被指数,如权重差值植被指数、土壤调整植被指数和转化土壤调整植被指数等等。这些植被指数不同程度上消弱了土壤背景噪声,但是必须预先已知下垫面植被的密度分布或覆盖百分比,也就是说仅适合于求解某一小范围植被覆盖变化较小下垫面上的植被指数,而且其动态范围也偏小。本文提出了一种修正的土壤调整植被指数,土壤调整因子由植被指数自身调整,不需要其它辅助信息。计算结果表明其动态范围较大,与最佳土壤调整因子下的土壤调整植被指数差异小,因此对应用卫星遥感资料求解区域面上土壤调整植被指数极其有效。     

Determining Effects of Area Burned and Fire Severity on Carbon Cycling and Emissions in Siberia

The Russian boreal forest contains about 25% of the global terrestrial biomass, and even a higher percentage of the carbon stored in litter and soils. Fire burns large areas annually, much of it in low-severity surface fires – but data on fire area and impacts or extent of varying fire severity are poor. Changes in land use, cover, and disturbance patterns such as those predicted by global climate change models, have the potential to greatly alter current fire regimes in boreal forests and to significantly impact global carbon budgets. The extent and global importance of fires in the boreal zone have often been greatly underestimated. For the 1998 fire season we estimate from remote sensing data that about 13.3 million ha burned in Siberia. This is about 5 times higher than estimates from the Russian Aerial Forest Protection Service (Avialesookhrana) for the same period. We estimate that fires in the Russian boreal forest in 1998 constituted some 14–20% of average annual global carbon emissions from forest fires. Average annual emissions from boreal zone forests may be equivalent to 23–39% of regional fossil fuel emissions in Canada and Russia, respectively. But the lack of accurate data and models introduces large potential errors into these estimates. Improved monitoring and understanding of the landscape extent and severity of fires and effects of fire on carbon storage, air chemistry, vegetation dynamics and structure, and forest health and productivity are essential to provide inputs into global and regional models of carbon cycling and atmospheric chemistry.     

用多光谱卫星信息分析白昼云天条件下的湿度场

TOVS资料的晴空湿度场反演已进行多年 ,如何利用卫星遥感信息实现云天条件下的湿度场反演 ,仍是一个值得研究的问题。多光谱卫星图像上不同云类的云层高低分布蕴涵着三维湿度场的分布信息 ,文中用相关分析、最小二乘拟合分析和多元线性回归分析分别讨论了多光谱卫星信息与各标准层相对湿度的关系。标准回归分析的复相关系数R在地面最小 ,约为 0 6 4,随高度增加 ,R逐层增大 ,70 0hPa以上大于 0 7,2 5 0hPa以上大于 0 8。相应显著性检验的结果F值从地面的 30 71,也逐层增大 ,至 2 0 0hPa时达 119 36。逐步回归分析在剔除了不显著的变量因子后 ,回归自变量减少到 2～ 4个 ,反映显著水平的F值有较大提高。在此基础上建立了一组云天条件下的多元线性回归方程 ,由此可得到具有卫星图像像素分辨率 (8～ 10km)的 9层湿度场。     

A Local-Scale Study of the Trace Gas Emissions from Vegetation Burning around the Village of Dalun, Ghana, with Respect to Seasonal Vegetation Changes and Burning Practices

The annual trace gas emissions from a West African rural region were calculated using direct observations of gas emissions and burning practices, and the findings compared to the guidelines published by the IPCC. This local-scale study was conducted around the village of Dalun in the Northern Region of Ghana, near the regional capital of Tamale. Two types of fires were found in the region – agricultural fires andwildfires. Agricultural fires are intentionally set in order to remove shrub and crop residues; wildfires are mostly ignited by herders to remove inedible grasses and to promote the growth of fresh grass. An agricultural fire is ignited with a fire front moving against the wind (backfire), whereas a wildfire moves with the wind (headfire). Gas emissions (CO₂, CO and NO) weremeasured by burning eight experimental plots, simulating both headfires and backfires. A common method of evaluating burning conditions is to calculate modified combustion efficiency (MCE), which expresses the percentage of the trace gases released as CO₂. Modified combustion efficiency was95% in the wildfires burned as headfires, but only 90% in the backfires.The burned area in the study region was determined by classifying a SPOT HRV satellite image taken about two months into the dry season. Fires were classified as either old burned areas or new burned areas as determined by the gradient in moisture content in the vegetation from the onset of the dry season. Classified burned areas were subsequently divided into two classes depending on whether the location was in the cultivated area or in the rangeland area, this sub-classification thus indicating whether the fire had been burned as a backfire or headfire. Findings showed that the burned area was 48% of the total region, and that the ratio of lowland wildfiresto agricultural fires was 3:1. The net trace gas release from the classified vegetation burnings were extrapolated to 26–46×10⁸ gCO₂, 78–302×10⁶ g CO,17–156×10⁵ g CH₄,16–168×10⁵ g NMHC and 11–72×10³ NO_x. Calculation of the emissionsusing proposed IPCC default values on burned area and average biomass resulted in a net emission 5 to 10 times higher than the measured emission values. It was found that the main reason for this discrepancy was not the emission factorsused by the IPCC, but an exaggerated fuel load estimate.     

Mangrove conversion and aquaculture development in Vietnam: A remote sensing-based approach for evaluating the Ramsar Convention on Wetlands

Remote sensing data have been proposed as a potential tool for monitoring environmental treaties. However, to date, satellite images have been used primarily for visualization, but not for systematic monitoring of treaty compliance. In this paper, we present a methodology to operationalize the use of satellite imagery to assess the impact of the Ramsar Convention on Wetlands. The approach uses time series analysis of landscape pattern metrics to assess land cover conditions before and after designation of Ramsar status to monitor compliance with the Convention. We apply the methodology to two case studies in Vietnam and evaluate the success of Ramsar using four metrics: (1) total mangrove extent; (2) mangrove fragmentation; (3) mangrove density; and (4) aquaculture extent. Results indicate that the Ramsar Convention did not slow the development of aquaculture in the region, but total mangrove extent has remained relatively constant, primarily due to replanting efforts. Yet despite these restoration efforts, the mangroves have become fragmented and survival rates for replanting efforts are low. The methodology is cost effective and especially useful to evaluate Ramsar sites that rely mainly on self-reporting methods and where third parties are not actively involved in the monitoring process. Finally, the case study presented in this paper demonstrates that with the appropriate satellite record, in situ measurements and field observations, remote sensing is a promising technology that can help monitor compliance with international environmental agreements.     

C波段高山天气雷达对森林火灾的探测能力初析

借鉴前人在S波段新一代天气雷达上探测森林火灾和城市大火烟尘的方法和研究成果,利用收集到的云南十部C波段新一代天气雷达的34次森林火灾个例资料,立足云南C波段高原高山雷达的一些特点,采用分类对比进行分析研究,得出C波段高山雷达森林火灾回波的一些特有指标。提出按雷达海拔高度分别设置森林火灾判别阈值、引入速度场特征、降水回波过滤等一些新指标和方法,进行这类地区的雷达森林火灾探测。随后开展了业务实践,经检验和实地踏勘,取得了一定效果。证明C波段雷达同样具有一定的探测森林火灾回波的能力,主要技术指标有:高山雷达由于海拔高差的影响,应分别针对不同雷达设置相应的阈值参数;根据森林火灾回波的上升和飘散特性,雷达回波速度场与周围回波有时呈现出一定差别;采用过滤晴空回波、过滤杂波、过滤地物回波、过滤二次回波、过滤降水回波、识别速度场特殊回波等办法,可得到滤出的火灾烟尘疑似回波。这些指标和方法对西部高山雷达探测应有一些参考作用。     

Climate change adaptation of coffee production in space and time

Savannas constitute the most fire-prone vegetation type on earth and are a significant source of greenhouse gas emissions. Most savanna fires are lit by people for a variety of livelihood applications. ‘Savanna burning’ is an accountable activity under the Kyoto Protocol, but only Australia, as a developed economy, accounts for emissions from this source in its national accounts. Over the past decade considerable effort has been given to developing savanna burning projects in northern Australia, combining customary indigenous (Aboriginal) approaches to landscape-scale fire management with development of scientifically robust emissions accounting methodologies. Formal acceptance by the Australian Government of that methodology, and its inclusion in Australia’s developing emissions trading scheme, paves the way for Aboriginal people to commercially benefit from savanna burning projects. The paper first describes this Australian experience, and then explores options for implementing community-based savanna burning emissions reduction projects in other continental savanna settings, specifically in Namibia and Venezuela. These latter examples illustrate that savanna fire management approaches potentially have broader application for contributing to livelihood opportunities in other fire-prone savanna regions.     

湖南省森林火灾时空分布特征及其气象成因分析

利用湖南省2008-2017年10年森林火灾数据、卫星MODIS和VIIRS监测资料及常规气象资料，统计分析了湖南省森林火灾时空分布特征，分析结果表明：近10年湖南省森林火灾高发月份是2、3、4月， 3月最多，总次数为3800次，森林火灾在邵阳发生次数最多，郴州、长沙和永州次之。森林火灾受灾面积最大的是永州市、怀化市、邵阳市，达6800-7500公顷。重点分析了2018年2月14-19日湖南致灾严重的森林火灾过程与气象条件的关系，结果表明，森林火灾期间，气温偏高、相对湿度较小，降水量少，平均风速一般超过0.5m/s，火灾严重时段最大风速增大到6m/s左右，风向以偏东风和偏南风为主，湖南高低空受上下一致的西南风气流控制，天气形势较稳定。     

利用静止气象卫星红外通道遥感监测中国沙尘暴

气象卫星的红外窗区通道 (8~12 μm) 对于通常大气气溶胶几乎没有响应, 但对于较大颗粒且浓度较强的沙尘气溶胶, 尤其是沙尘暴有明显的信号反应。空气中的沙尘在红外分裂窗通道表现出两个特征:一是对地表发射到空间的红外信号产生衰减, 造成卫星探测到的地气系统亮温降低, 这就是所谓的红外差值沙尘指数IDDI; 二是沙尘粒子在红外分裂窗两个通道的比辐射率不同, 11 μm比12 μm的比辐射率低, 从而造成这两个通道的亮温差是负值。基于这两个特征和沙尘多通道光谱聚类法, 针对静止气象卫星观测数据进行了沙尘暴卫星遥感监测业务算法开发, 输出沙尘暴监测产品和红外差值沙尘指数产品, 这一算法不仅用于已经退役的GMS-5卫星, 而且应用于正在运行的静止气象卫星FY-2C, 它还为沙尘暴的定量或半定量遥感提供参考借鉴。     

卫星遥感监测森林火情传输服务系统

卫星遥感监测森有传输服务系统是为用户建立的服务界面，依托于计算机网络通讯，用微机把卫星遥感森林火情监测信息展示在用户面前，彻底改变了以前只能用电话传输火情的信息方法，使用户能快速，高效和全面了解火点信息，为防火指挥机构及时掌握火情动态和指挥扑救火灾提供可靠依据。     

基于遥感和地理信息系统的河南省森林防火系统

河南森林火灾较多,给森林资源造成较大损失。由于在火灾预警、监测、救灾中,没有一套有效的现化化的手段,常常延误较多的扑救时间,造成不必要的经济和森林资源的损失。在多年森林火灾遥感监测的基础上,利用“3S”技术,从森林火灾预警—监测—扑救等过程入手,建立了自动化、流程化的河南省森林遥感防火系统。介绍了河南省森林遥感防火系统的主要技术思路及各部分主要功能。该系统的资料接收、处理、服务等都实现了软件支持,并利用“3S”技术,使火点监测图像、气象和森林火险预报与地理信息叠加。     

梧州市森林火险气象预报

使用１９８０－１９９１年森林火灾及相应的气象资料，分析了森林火灾与有关气象要素的相关性。在此基础上，计算了森林火险天气预报指标，建立了预报模式，开展了森林火险天气等级预报。经试用效果较好。     

气候变化对林火动态的影响研究进展

气候变化对林火动态的影响一直是国内外研究的热点,从数据获取方式、研究内容、研究方法和研究结论等角度出发对此领域的进展进行了综述。当前以气候变暖为主要特征的气候变化已对林火动态产生重要影响,全球大部分地区林火发生频率增大、林火强度增强;在预期继续变暖的气候背景下,未来林火情势会更加严峻;但由于气候变化的区域性差异,有些地区的林火动态变化不明显或呈现下降趋势。对预期气候情景下研究结论的不确定性进行了分析,并讨论了我国与国外在此领域的研究差距及其原因。最后,对未来的研究方向进行了展望。