基于遗传算法优化BP神经网络的玉米遥感估产方法     

于海洋  陈圣波  杨北萍  安秦 《世界地质》2020,39(1):208-214

以内蒙古自治区开鲁县玉米作物为研究对象,将生育期内玉米遥感影像所提取的多种植被指数和实地采样点的测产数据作为训练值,利用BP(back propagation)神经网络和遗传算法优化BP(GA-BP)神经网络估产模型,得出网络预测的玉米产量数值。通过决定系数R 2和均方根误差RMSE,比较实测产量与预测产量之间的精度,BP神经网络模型R^2为0.8452,RMSE(%)为28.37;遗传算法优化BP神经网络模型R^2为0.9850,RMSE(%)为6.70,表明遗传算法优化BP神经网络估产模型具有一定可行性和可信度。  相似文献   

南方亚热带季风区将乐县森林植被动态变化及其对气候变化的响应     

左冰洁  孙玉军 《地球信息科学学报》2019,21(6):958-968

在全球气候变化背景下,植被动态变化以及植被对气候变化的响应方式已经成为生态学和地理学领域的热点。本文对比分析了南方亚热带季风区将乐县不同类型森林植被对不同时间尺度的干旱响应的差别。基于2000-2017年MODIS-EVI数据及气象站点数据,用最大值合成法、趋势分析法以及相关分析法,分析了森林植被及气象因子的动态变化特征,并对比不同森林植被对气候变化响应的差别。研究表明：① 2000-2017年,研究区植被覆盖度、EVI和降水均显著增加,区域内湿度增加,森林长势渐趋良好;② EVI在生长季初期和末期与同期的降水、温度均显著正相关（P<0.1）,初期森林受降水因子的影响更大,末期受温度因子的影响大;③ 1-3月和周年的气候变化对森林的生长至关重要,长时间尺度的湿度增加对森林生长具有显著的促进作用,SPEI的时间尺度越长与EVI的相关性也越大;④ 针阔混交林与同期温度、降水的相关系数最高,并且与不同时间尺度的SPEI相关性均比较高,属于气候敏感型林型,在生产经营中要谨慎预防气候变化对该林型带来的伤害;⑤ 森林覆盖度变化与降水和SPEI_24的相关性极显著,长时间尺度的降水变化是影响森林植被覆盖率变化的重要因素之一。  相似文献   

Acceleration of vegetation dynamics in hydrologically connected wetlands caused by dam operation     

Lilin Zheng  Jinying Xu  Dandan Wang  Guoyu Xu  Zhiqiang Tan  Ligang Xu  Xiaolong Wang 《水文研究》2021,35(2):e14026

The dynamic responses of wetlands to upstream water conservancy projects are becoming increasingly crucial for watershed management. Poyang Lake is a dynamic wetland system of critical ecological importance and connected with the Yangtze river. However, in the context of disturbed water regime in Poyang Lake resulting from human activities and climate change, the responses of vegetation dynamics to the Three Gorges Dam (TGD) have not been investigated. We addressed this knowledge gap by using daily water level data and Landsat images from 1987 to 2018. Landsat images were acquired between October and December to ensure similar phenological conditions. Object-oriented Artificial Neural Network Regression for wetland classification was developed based on abundant training and validation samples. Interactions between vegetation coverage and water regimes pre and post the operation of the TGD were compared using classification and regression trees and the random forest model. Since the implementation of the TGD in 2003, Poyang Lake has become drier, especially during the dry season. A more rapid plant growth rate was observed post TGD (44.74 km² year⁻¹) compared to that of the entire study period (12.9 km² year⁻¹). Average water level for the antecedent 20 days most significantly affected vegetation before 2003, whereas average water level for the antecedent 5 or 10 days was more important after 2003. The impoundment of the TGD after the flood season accelerated the drawdown processes of Poyang Lake, and the rapidly exposed wetlands accelerated vegetation expansion during the dry seasons, resulting in shrinkage and degradation of the lake area. This study deepens our knowledge of the influences of newly developed dams on lakes and rivers.  相似文献   

1996—2015年黄河源区植被覆盖度提取和时空变化分析   总被引：2，自引：2，他引：0  

王俊奇  王广军  梁四海  杜海波  彭红明 《冰川冻土》2021,43(2):662-674

研究黄河源区植被覆盖度时空变化对于深入理解青藏高原多年冻土区在气候变化和人类活动双重作用下的植被响应,以及为黄河源区生态环境保护和治理提供决策具有重要的意义。以陆地卫星（Landsat）影像为主要数据源,利用多端元混合像元分解模型（Multiple Endmember Spectral Mixture Analysis,MESMA）,完成了1996—2015年黄河源区4.4万km²、7个时相的植被覆盖度提取,并基于转移矩阵和一元线性回归趋势法分析了植被覆盖度变化情况。结果表明：黄河源区东南部植被覆盖度较高,西北部植被覆盖度较低,且植被覆盖度在空间上由东南向西北呈递减趋势。1996—2004年植被覆盖度整体呈下降趋势,2004—2015年植被覆盖度呈增加趋势。1996—2015年植被覆盖度呈增加趋势的区域占57.25%,基本不变的区域占16.02%,植被覆盖度呈下降趋势的区域占26.73%。植被覆盖度下降的主要原因是黄河源头及一些河谷地带、环湖地区受人类影响较大,且东南部海拔较低地区受到过度放牧的影响。尽管黄河源区1996—2015年植被覆盖度总体呈改善趋势,但毒杂草的面积也由1996年的16 060 km²增加到2015年的22 942 km²,20年间增加了6 882 km²,毒杂草面积的增加对黄河源区局部地区畜牧业的发展有不利影响。  相似文献   

Late Pleistocene and Holocene Afromontane vegetation and headwater wetland dynamics within the Eastern Mau Forest,Kenya     

Esther N. Githumbi  Colin J. Courtney Mustaphi  Rob Marchant 《第四纪科学杂志》2021,36(2):239-254

The Mau Forest Complex is Kenya's largest fragment of Afromontane forest, providing critical ecosystem services, and has been subject to intense land use changes since colonial times. It forms the upper catchment of rivers that drain into major drainage networks, thus supporting the livelihoods of millions of Kenyans and providing important wildlife areas. We present the results of a sedimentological and palynological analysis of a Late Pleistocene–Holocene sediment record of Afromontane forest change from Nyabuiyabui wetland in the Eastern Mau Forest, a highland region that has received limited geological characterization and palaeoecological study. Sedimentology, pollen, charcoal, X-ray fluorescence and radiocarbon data record environmental and ecosystem change over the last ~16 000 cal a bp. The pollen record suggests Afromontane forests characterized the end of the Late Pleistocene to the Holocene with dominant taxa changing from Apodytes, Celtis, Dracaena, Hagenia and Podocarpus to Cordia, Croton, Ficus, Juniperus and Olea. The Late Holocene is characterized by a more open Afromontane forest with increased grass and herbaceous cover. Continuous Poaceae, Cyperaceae and Juncaceae vegetation currently cover the wetland and the water level has been decreasing over the recent past. Intensive agroforestry since the 1920s has reduced Afromontane forest cover as introduced taxa have increased (Pinus, Cupressus and Eucalyptus).  相似文献   

黄土高原北部土地利用变化对长期土壤水分平衡影响模拟   总被引：3，自引：0，他引：3       下载免费PDF全文

白晓  贾小旭  邵明安  赵春雷 《水科学进展》2021,32(1):109-119

为探明土地利用变化对黄土高原长期土壤水分平衡的影响,利用校验的Hydrus-1D模型模拟黄土高原北部神木六道沟小流域1981—2050年农耕地—苜蓿草地—天然草地情景下0~4 m土壤水分变化过程,量化土壤储水量、深层渗漏和蒸散发等水文变量的演变特征。结果显示:①农耕地期间年降水的88%为蒸散发消耗,11%为渗漏损失。②苜蓿草地种植后6 a内,蒸散发大幅增加至年降水的108%,土壤水分负平衡,0~4 m土壤储水量以52 mm/a的速率降低;至7~13 a,年降水几乎全部被蒸散发消耗。③苜蓿草地转变为天然草地后,蒸散发量下降31%,土壤水分以45 mm/a的速率逐渐补给,之后年降水量的92%用于蒸散发,8%为渗漏消耗,土壤水分处于相对稳定状态。研究表明不同土地利用方式下的土壤水分平衡模式具有显著差异,种植高耗水植被可造成土壤水分负平衡,导致土壤干燥化,进而对土壤水分补给产生负面效应,改变植被类型可使土壤干层得到有效改善。  相似文献   

The relationship between groundwater and natural vegetation in Qaidam Basin          下载免费PDF全文

Dang Xue-ya  Lu Na  Gu Xiao-fan  Jin Xiao-mei 《地下水科学与工程》2021,9(4):341-349

To accurately evaluate ecological risks trigged by groundwater exploitation, it must be clarified the relationship between vegetation and groundwater. Based on remote sensing data sets MOD13Q1, groundwater table depth (WTD) and total dissolved solids (TDS), the relationship between groundwater and natural vegetation was analyzed statistically in the main plain areas of Qaidam Basin. The results indicate that natural vegetation is groundwater-dependent in areas where WTD is less than 5.5 m and TDS is less than 7.5 g/L. Aquatic vegetation, hygrophytic vegetation and hygrophytic saline-alkali tolerant vegetation are mainly distributed in areas with WTD <1.1 m. Salt-tolerant and mesophytic vegetation mainly occur in areas with WTD of 1.4-3.5 m, while the xerophytic vegetation isprimarily present in areas where WTD ranges from 1.4 m to 5.5 m. Natural vegetation does not necessarily depend on groundwater in areas with WTD >5.5 m. For natural vegetation, the most suitable water TDS is less than 1.5 g/L, the moderately suitable TDS is 1.5-5.0 g/L, the basically suitable TDS is 5.0-7.5 g/L, and the unsuitable TDS is more than 7.5 g/L.  相似文献   

基于小波的浙江省NDVI与自然—人文因子多尺度空间关联分析   总被引：1，自引：0，他引：1  

徐芝英  胡云锋  甄霖  庄大方 《地理研究》2015,34(3)

对地表植被的空间分布格局及其与自然、人文因子间的关系开展综合的、多尺度的定量分析,这是全球气候变化背景下土地利用与覆被研究中的一项重要内容。以地处亚热带湿润季风区的浙江省为研究区,设置了东西向(样线A)和北南向(样线B)两条样线,并应用小波分析方法对研究区NDVI、高程、坡度和土地利用强度等因子的尺度特征以及这些因子间的多尺度相关关系进行了分析。结果表明:1研究区内,NDVI、高程、坡度及土地利用强度的空间分布格局存在四个尺度域;2上述因子在样线A及样线B上的空间分布分别存在两个主要特征尺度,其中东西向(样线A)上的特征尺度为40 km和80 km,北南向(样线B)上主要的特征尺度为30 km和50 km;3在较大尺度域(8 km以上),土地利用强度是影响NDVI空间分布的最主要因素,而在小尺度域(0~8 km),坡度和高程因子成为影响NDVI分布的主要因子。研究还认为,小波分析方法为识别地理要素空间分布的特征尺度、量测要素间任意尺度、任意位置上的相干关系,提供了方便的工具。  相似文献   

对地观测高光谱激光雷达发展及展望     

龚威  史硕  陈必武  宋沙磊  牛铮  王成  管海燕  李伟  高帅  林沂  孙嘉  杨健  杜霖 《遥感学报》2021,25(1):501-513

对目标空间三维—光谱信息的高分辨一体化获取与应用,是对地观测技术发展的前沿科学问题。结合高光谱成像与激光雷达测距的技术优势,对地观测多光谱/高光谱激光雷达遥感技术手段应运而生,并成为遥感技术未来发展的重要方向。本文分3个阶段详细回顾了对地观测高光谱激光雷达系统的发展历程,并针对其独有数据类型阐述了数据处理研究方面的探索研究。最后,重点分析了高光谱激光雷达在测绘领域、农林业领域的重大应用潜力,展望了未来对地观测高光谱激光雷达发展面临的机遇和挑战。  相似文献   

植被生态遥感参数定量反演研究方法进展     

赵燕红  侯鹏  蒋金豹  姜赟  张兵  白君君  徐海涛 《遥感学报》2021,25(11):2173-2197

植被参数是生态遥感定量反演的热点和难点,也是生态系统研究的基础性参数。本文在广泛阅读国内外公开发表的文献资料基础上,将现有的植被生态遥感参数概括为物理类、生化组分类、能量和功能类3大类,系统梳理了每类参数定量反演的主要模型方法,进行优缺点和适用性分析,对现阶段存在的不足和未来发展趋势进行了探讨。物理类植被生态遥感参数主要介绍了植被覆盖度、生物量、叶面积指数、树高等研究进展,生化组分类植被生态遥感参数主要介绍了植被含水量、叶绿素含量和光合能力等研究进展,能量类植被生态遥感参数介绍了光合有效辐射、植被吸收光合有效辐射等研究进展,功能类植被生态遥感参数主要介绍了植被生产力和碳交换量等研究进展。存在的主要问题包括混合像元分解、病态反演、物理模型应用过程中的误差传递、数据融合的尺度效应和空间变异性、模型耦合的最优方案确定问题。  相似文献