Fire is a worldwide phenomenon that appears in the geological record soon after the appearance of terrestrial plants. Fire influences global ecosystem patterns and processes, including vegetation distribution and structure, the carbon cycle, and climate. Although humans and fire have always coexisted, our capacity to manage fire remains imperfect and may become more difficult in the future as climate change alters fire regimes. This risk is difficult to assess, however, because fires are still poorly represented in global models. Here, we discuss some of the most important issues involved in developing a better understanding of the role of fire in the Earth system.

【目的】 利用森林火灾风险图可提高有效巡护,优化有限防火资源,基于地形、人类活动、植被和气象因素数据,采用基于机器学习算法构建了林火发生预测模型,对林火防护提供一定的参考。【方法】 以安徽省滁州韭山为研究对象,提取林区的坡度、海拔、坡向、到居住点的距离、到道路的距离、地形湿度指数、归一化植被指数和温度驱动因素,评估火灾发生驱动因子,将潜在驱动因子分成地形、人类活动、植被与气象因素等4类;使用哨兵火灾产品,提取林区内的历史火点,然后采用机器学习算法建立林火发生的预测模型;最后利用混淆矩阵评估指标和接受者操作特征曲线(ROC)进行精度评价。【结果】 植被、温度和到道路的距离是研究区域火灾发生的主要驱动因素。两种模型的ROC曲线表明,逻辑回归预测模型准确度为71.07%,曲线下面积(AUC)值为0.717 2;随机森林模型具有较好的准确性,准确度达到84.91%,曲线下面积值为0.850 1。【结论】 随机森林模型表现出比逻辑回归模型更好的预测能力。森林火灾风险图表明,随机森林模型预测下,研究区11.91% (29.36 km²)位于高、极高风险级别。森林火灾风险图可有效协助林火防护管理者采取适当的林火预防措施,保护森林资源。

燃料载荷和含水率是森林火灾中地表火蔓延的重要影响因素。选取香樟叶和云南松松针作为研究对象，开展了不同燃料载荷和含水率条件下平坡地表火蔓延实验。研究结果表明：同一燃料载荷下，随含水率增大，火焰长度、火蔓延速率、火线强度和辐射热流减小，滞留时间增大；同一含水率下，燃料载荷增大，火焰长度、火蔓延速率、火线强度、辐射热流和滞留时间均增大。

Fuel load and water content are important factors affecting surface fire spread in forest fires.This work selected cinnamomumcamphora leaf and pine needle of pinusyunnanensis as the research object，carried out the fire spread experiment under flat slope conditions，with different fuel load and moisture content.It was observed that under the same fuel load condition，the flame length，the rate of spread，the fireline intensity and the radiative heat flux decreased with the increase of water content，while the residence time increased.Under the same moisture content condition，the flame length，the rate of spread，the fireline intensity，the radiative heat flux and the residence time increased with increasing fuel load.

【目的】建立快速、无损监测棉花冠层等效水厚度（canopy equivalent water thickness,CEWT）的估算模型,进一步提高利用高光谱遥感技术监测棉花CEWT的估算精度。【方法】通过在不同生育期设置灌溉梯度试验,于棉花蕾期和花铃期同步测定冠层光谱反射率、冠层等效水厚度等信息,综合分析棉花冠层等效水厚度与原始光谱反射率、一阶导数光谱反射率、全波段组合光谱指数、已有光谱指数的相关性,确定蕾期、花铃期及全生育期的最优光谱指数,并通过线性回归构建棉花CEWT的高光谱监测模型。【结果】冠层等效水厚度与原始光谱反射率在近红外波段（NIR）780—1 130 nm和短波红外波段（SWIR）1 450—1 830 nm、1 950—2 450 nm附近均出现连续的敏感波段,一阶导数光谱在NIR波段内对CEWT的敏感程度较原始光谱有所加强,但在SWIR波段内敏感程度弱于原始光谱;利用原始光谱反射率构建的光谱指数与CEWT的相关性强于一阶导数光谱,且比值光谱指数（RSI）较归一化差分光谱指数（NDSI）更适合CEWT的监测。在全生育期内(R₁₁₃₅-5R₁₄₉₄)/R₂₀₀₃对CEWT的反演精度最佳（R ²=0.7878,RRMSE=0.1803）;在蕾期RSI_b(1130,1996)对CEWT的估算效果最好（R ²=0.7258,RRMSE=0.1444）;在花铃期RSI_a（904,1952）是估算CEWT的最优光谱指数（R ²=0.7853,RRMSE=0.2454）。【结论】该研究在不同生育阶段内提出的新型高光谱指数均可用于棉花冠层等效水厚度的定量监测,研究结果可为高光谱技术在棉花冠层水分含量监测中的应用提供参考,为棉花精准灌溉提供技术依据。

【Objective】 The objective of the experiments is to develop a key method for fast and nondestructive monitoring canopy equivalent water thickness (CEWT) in cotton (Lumian 54) and to further improve the estimation accuracy of CEWT in cotton monitored by remote sensing technology. 【Method】 Through setting irrigation gradient treatment in different growth period, canopy spectral reflectance and canopy equivalent water thickness and other information were measured simultaneously. Firstly, we comprehensively analyzed the correlation between CEWT and various spectral parameters, including original spectral reflectance, first derivative spectral reflectance, all-band combined spectral index and existing spectral index. Then, we determined the optimal spectral indices of bud stage, flowering and bolls stage, and full growth period. Finally, we constructed a hyperspectral monitoring model of cotton CEWT by linear regression. 【Result】 The canopy equivalent water thickness and the original spectral reflectance show continuous sensitive bands in the near infrared band (NIR) of 780-1130 nm and the short wave infrared band (SWIR) of 1 450-1 830 nm and 1 950-2 450 nm, the sensitivity of the first derivative spectrum to CEWT was enhanced in NIR band than that of the original spectrum, but was weaker in SWIR band than that of the original spectrum. The correlation between the spectral index constructed by the original spectral reflectance and CEWT is stronger than that of the first derivative spectrum, and the ratio spectral index (RSI) is more suitable for the monitoring of CEWT than the normalized difference spectral index (NDSI). During the whole growth period, the inversion accuracy of CEWT by (R₁₁₃₅-5R₁₄₉₄)/R₂₀₀₃ was the best (R ²=0.7878, RRMSE=0.1803). In the bud stage, RSI_b(1130,1996) has the best estimation effect on CEWT (R ²=0.7258, RRMSE=0.1444). RSI_a (904,1952) was the optimal spectral index (R ²=0.7853, RRMSE=0.2454) for estimating CEWT at the flowering and bolls stage.【Conclusion】The new hyperspectral indexes proposed in this study in different growth stages can be used for quantitative monitoring of canopy equivalent water thickness in cotton. The results of this study can provide reference for the application of hyperspectral technology in monitoring water content of cotton canopy, and provide technical basis for precision irrigation of cotton.

叶片等效水厚度(EWT)是评估油菜生长状态的一个重要参数。为快速准确估算油菜叶片EWT,选择9个常用的植被水分指数(WI、PRI、NDVI、NDII、NDWI、MSI、PWI、GVMI、NDMI),在6个已有角度指数(β_SWIR₁、SANI、SASI、ANIR、NANI、NASI)基础上,提出2种角度比值指数(SARI、NARI),并根据油菜叶片水吸收谷峰高光谱特征,提出基于水吸收谷1 450 nm和1 930 nm的8种改进型角度指数,利用以上25种角度指数估算不同施氮水平下苗期、蕾薹期以及不区分苗期、蕾薹期情况下的油菜叶片EWT。结果表明,苗期ANI₁₄₅₀、ASI₁₄₅₀、MSI、GVMI、NDII估算效果较好,R²均达到0.81以上;蕾薹期ANI₁₉₃₀、ASI₁₉₃₀、NASI、SANI、GVMI、SARI效果最好,R²均达到0.71以上;在不区分苗期、蕾薹期的情况下,改进型角度指数ANI₁₄₅₀、ASI₁₄₅₀效果最好,R²均达到0.832,可以在不区分苗期、蕾薹期情况下对油菜叶片EWT进行估算,适用性更广。本研究提出的改进型角度指数不仅丰富了已有角度指数,且提高了其反演油菜叶片EWT的精度,为快速精确估计油菜叶片EWT提供了新的研究思路。

Leaf equivalent water thickness (EWT) is an important parameter to evaluate the growth state of rapeseed. To estimate EWT of rapeseed leaves quickly and accurately, 9 commonly used vegetation water indices (WI, PRI, NDVI, NDII, NDWI, MSI, PWI, GVMI, NDMI) are selected, and two kinds of angle ratio indices (SARI, NARI) are proposed based on the six existing angle indices (&#x003b2;_SWIR1, SANI, SASI, ANIR, NANI, NASI). According to the hyperspectral characteristics of the water absorption valleys and peaks of rapeseed, 8 improved angle indices are brought forward based on water absorption valley at the wavelengths of 1450 nm and 1930 nm. The above 25 angel indices are used to estimate leaf EWT of rapeseed under different nitrogen levels at the seeding stage, bud stage and without distinguishing between seedling stage and bud stage. The results show that ANI₁₄₅₀, ASI₁₄₅₀, MSI, GVMI and NDII performed better at the seeding stage, and their <em>R</em>² both achieved 0.81 or above. ANI₁₉₃₀, ASI₁₉₃₀, NASI, SANI, GVMI and SARI are the best at the bud stage as their <em>R</em>² are all larger than 0.71. Improved angle indices ANI₁₄₅₀ and ASI₁₄₅₀ have the best effect, with their <em>R</em>² at 0.832, which can be used to estimate the EWT of rapeseed leaves without distinguishing between seedling stage and bud stage, and the applicability is wider. The improved angle indices in this research can not only enrich the existing angle indices, but also improve the accuracy of rapeseed leaf EWT inversion. Moreover, the indices provide a new idea for estimation of rapeseed leaves EWT instantly and precisely.

基于PROSAIL模型，结合野外实测叶片等水分厚度、干物质重量、叶面积指数数据，构建一种基于归一化红外指数和归一化干物质指数的植被冠层可燃物含水率估算方法。首先，在PROSAIL模型输入实测参数模拟植被冠层光谱曲线，计算归一化红外指数、归一化干物质指数用于叶片等水分厚度、干物质重量的反演。结果表明：归一化红外指数与叶片等水分厚度、归一化干物质指数与干物质重量存在明显的线性关系，基于该关系建立叶片等水分厚度、干物质重量的经验估算模型，经验证估算结果精度较高；将该经验模型推广至利用Landsat 8数据拟合植被冠层可燃物含水率，并与实测数据进行验证，结果显示R2达到0.743，RMSE达到34.2%，具有较高的精度。文章提出的植被冠层可燃物含水率估算模型，可实现广州市过渡带森林大面积、较高精度植被冠层可燃物含水率监测，为预防森林火灾提供参考。

Fuel moisture content (FMC), which is the ratio of equivalent water thickness (EWT) to dry matter content (DMC), plays a crucial role in the estimation of vegetation ignition probability and the fire propagation rate. The PROSAIL model can adequately simulate the canopy reflectance of vegetation, with the input of field-measured data into the model ensuring conformity with the ecological rules. If the EWT and DMC can be estimated by an empirical statistical method according to the reflectance spectrum, the universality of the physical model and the efficiency of the empirical statistical method can be considered. In this study, a fast and versatile method is established for estimating FMC based on the EWT, DMC, leaf area index measured data, and the PROSAIL model. The Normalized Difference Infrared Index (NDII) and Normalized Dry Matter Index (NDMI) were obtained from the spectral curves, with the results showing an obvious linear relationship between the NDII and EWT, NDMI, and DMC. Therefore, EWT and DMC can be estimated using the NDII and NDMI. The accuracy of the estimation results is verified to be high. The estimation model can be extended to Landsat 8 data to estimate FMC. The estimated FMC data verified by the measured data showed that R2 was 0.743 and the RMSE was 34.2%. The model accuracy was reliable owing to large dynamic changes in the FMC. However, the estimated value of the FMC shifted significantly to the left during this study. The reasons for this are as follows: 1) The field-measured points are not sufficient to support the analysis according to different vegetation types, and the physical and chemical properties of different types are varied, leading to altered simulated spectral curves; 2) The vegetation spectrum is sensitive to the DMC at 1,650 nm, 1,720 nm, and 2,270 nm bands, and the sensitivity near the 1,650 nm and 1,720 nm bands is greater than that at 2,270 nm. However, because the Landsat 8 image does not have a 1,720 nm band, the 2,270 nm band was selected to calculate the NDMI, making the value of the estimated DMC too large, resulting in a small value of the estimated FMC and a significant shift to the left; 3) 1,650 nm and 2,270 nm are not in the central wavelength of the two bands of Landsat 8; therefore, the estimated DMC and FMC are shifted. In addition, the fast and versatile method, which is established based on the EWT, DMC, leaf area index measured data and the PROSAIL model, can realize large-scale and high-precision monitoring of the FMC, providing a scientific reference for forest fire prevention.

植被含水量(VWC)能够指示植被的水分状况,对植被生长、火灾、旱灾以及生态环境安全监测等具有重要意义,也是微波遥感估算土壤水分的重要参数之一.光谱指数法是估算植被含水量最常用的方法之一.结合地面观测及Landsat８OLI传感器遥感影像,对平凉地区的植被含水量进行了遥感估算模型研究,结果表明:①平凉地区叶片含水量(FMC)与植被光谱指数没有相关关系,而等效水深(EWT)则与各植被光谱指数具有显著的相关关系(均超过９５％显著性水平),其中RVI２与EWT的相关关系最显著且最稳定;②利用RVI２对研究区EWT进行遥感估算,其均方根误差(RMSE)为０．１８３,平均相对误差为８．９％,平均相对误差绝对值为２６．４％;③研究区内大部分农田的植被含水量为０．６~０．９kg/m２,少数农田的植被含水量达到１kg/m２ 以上,这与实际考查基本一致,基本能够反映研究区内农田EWT的空间变化特征.

为了筛选苏北丘陵区防火林带树种,以盱眙防火林带的建设为例,通过锥形量热仪对防火林带香樟、女贞、侧柏、广玉兰、杨树、海桐、油茶和淡竹等8种常见树种枝叶的16种指标进行了测定,并应用因子分析方法进行统计分析,得出了8种树种的抗火性差异,其由强到弱依次为广玉兰、杨树、女贞、油茶、侧柏、海桐、淡竹和香樟。