共查询到20条相似文献，搜索用时 468 毫秒
1.
2.
利用河西走廊伏旱和伏期降水资料序列,张掖观象台的地面气温、降水、探空等气象资料,以及国家气候中心提供的74个环流因子,借助BP神经网络可以逼近任意非线性函数的能力和特点,构建了一个用于预测伏旱和伏期降水的模型,并对模型的预报效果进行验证。结果表明:BP神经网络模型能够对伏期干旱进行有效地预测,该预测模型对伏旱和伏期降水有比较理想的预报效果,伏旱预报历史拟合率高达97.6%、模型试报准确率为84.6%,伏期降水预测历史拟合率高达97.6%、模型试报准确率为76.9%,其性能指标符合实际要求,具有很好的实际应用价值。 相似文献
3.
门槛自回归模型,是香港中文大学汤家豪教授,为解决非线性周期变化的时间序列提出的一种动态分析模型.由于该模型拟合效果好,又能解决时间序列中的间断,跳跃现象.因此近期在供电、防震、供水和铁路运输预测方面、逐渐被利用.本文利用门槛自回归模型对黑龙江省伊春市1982年至1986年气温进行了预测,经过三年的实阶验证.效果较好. 相似文献
4.
本文在降水分区的基础上,对西北地区夏季降水进行预报。利用2007—2010年的6—8月T639资料和相应时段的实况资料,通过概率回归降水等级方案进行建模,对2011年6—8月进行了试报。结果表明:与单站建模预报相比,分区建模降水预报TS评分在各时效、各量级上均有提高,并且在空报和漏报上有较大减少,特别是中雨预报改善明显。分区建模比单站建模所选因子更丰富,利用了模式产品的有用信息,因此预报效果更好。分区建模降水预报与模式直接输出降水预报的对比分析表明:分区建模的降水预报效果优于模式直接输出降水预报,尤其小雨预报效果显著,中雨和大雨36 h和60 h预报的空报现象明显减少。 相似文献
5.
扎龙湿地降水变化非线性特征研究 总被引：1，自引：0，他引：1
由于趋势、周期和随机因素的综合影响,降水时间序列往往呈现复杂波动的非线性特征,而分离三种因素在降水时间序列中的贡献率是目前较难解决的问题,相应的研究对降水变化分析和预测具有重要意义。该文基于数理统计原理、极大熵谱原理和分形原理对扎龙湿地1951-2008年的年降水时间序列进行分析,结果表明:该地区存在2.4a和6.4a的降水短周期;拟合程度比较高的非线性回归模型说明该地降水具有循环波动趋势;盒维数揭示了周期、趋势和随机因素是影响扎龙湿地降水变化的主要因素,其贡献率分别为76.97%、16.78%和6.25%,其中周期是最主要的影响因素。 相似文献
6.
地下水资源评价中降水量的时间序列--马尔可夫模型 总被引：8，自引：1，他引：8
降水量的预报精度对以其为直接或间接补给源的地下水资源评价具有重要的影响，将时间序列方法与随机过程离散状态的马尔可夫链理论相结合，提出了时间序列－马尔可夫模型以预报大气降水量，模型根据降水序列的特征，采用一维非平衡时间序列模型进行预测，预测结果总体效果较好，但峰值点处误差较大，为了提高模型对波动性较大随机变量的预报精度，利用随机过程马尔可夫概率状态转移矩阵预报方法对其预测值进行二次拟合，实例计算表明，时间序列－马尔可夫模型预报效果良好，预报精度明显高于单一的时间序列模型精度，该结果拓宽了时间序列预报模型的应用范围，且对以大气降水为最终补给源的地下水资源评价具有重要的实用价值及理论意义。 相似文献
7.
基于时空窗口改进的时空加权回归分析——以湖北省黄石市住房价格为例 总被引：3，自引：3，他引：0
针对地理加权回归(GWR)模型不能有效处理样本数据空间自相关性这一问题,构造局部时空窗口统计量,尝试改进时空加权回归(GTWR)模型。定义多时空窗口的概念,给出其选取、计算和验证方法;计算时空窗口包含的各样本点的被解释变量平均值,与样本拟合点的被解释变量值的比值,作为新的解释变量,构建改进的时空加权回归(IGTWR)模型。以土地稀缺、多中心、资源型城市——湖北省黄石市为例,收集2007~2012年商品住宅成交价格1.93万个数据和398个楼栋样本点,选取小区等级、绿化率、楼栋总层数、容积率、距区域中心距离和销售年份6个解释变量,分别利用常规线性回归(OLS)、GWR、GTWR和IGTWR方法进行回归分析。模型结果表明：计算Moran’s I指数和分析时间序列的自相关性,能确定时空窗口的大小和数量的选取;IGTWR模型和各变量的回归统计均通过0.05的显著性水平检验,有关解释变量的系数估计值在空间分布上能合理解释;GWR拟合结果优于OLS,GTWR优于GWR,而IGTWR拟合精度最好。与GTWR模型分析相比, IGTWR模型R2从0.877提升到0.919,而AICc、残差方(RSS)和均方差(MSE)分别从6 226、49 996 201和354.427下降到6 206、32 327 472和284.969。案例研究表明：IGTWR能够表达一定时空范围的时空自相关特征,减小了估计误差,提高了回归拟合精度。 相似文献
8.
9.
逆延华南气温序列方法的探讨 总被引：1，自引：0，他引：1
本文用香港１８８４－１９８３年逐年１月平均气温代表冬温，把它的年际变化时间序列作为原样本进行了谱分析，从中提取有代表性的周期波，用谱波分析和方差分析法，对选出来的周期波延，迭加，拟合成新的冬温时间序列。用地方志的冻灾史料进行了检验，结果９５％以上的冻灾年与新序列的冷冬年对应；冬季应温多出现连续性冷冬年，史料和新序列也有很好的一致性。 相似文献
10.
以重建塔里木河中游年径流量的拟合位为原始数据,用维纳滤波和线性预测方法作了塔里木河中游50年的年径流量预测。前20年(1964-1983)作为试报,用以考察预报效果,后30年(1984-2013)为预报。利用年径流量预报值分析未来30年南疆的气候变化趋势。 相似文献
11.
基于国内现行的森林火险气象指数和单因子火险贡献度模型,以及逻辑回归模型和随机森林模型,在林火预报中引入微波遥感土壤水分信息,使用MCD14DL火点数据集和地面气象观测资料对广东省不同时间尺度的林火发生概率进行预测。结果表明：逻辑回归模型和随机森林模型构建的林火预测模型显著优于现行的森林火险气象指数和单因子火险贡献度模型,预测精度提升约20%。其中,随机森林模型对林火频数的解释程度最高（两者相关系数为0.476）。此外,加入微波土壤水分信息后,相较原有的基于气象要素的林火预测模型,2种机器学习模型的预测精度均略有提升,体现了表层土壤水分信息在林火预报中的重要性。研究可为高效提取对地观测信息,以改进华南地区不同时间尺度的林火预报工作提供参考。 相似文献
12.
基于GIS的农业气候资源区域化问题研究——以甘肃省为例 总被引：10，自引：1，他引：9
在农业气候资源研究中,站点数据的区域化问题是进行资源优化配置和高效利用的一个重要环节。通过采用逐步回归分析与空间插值相结合的方法,以甘肃省及其相邻省区的112个站点1970~2001年31年的月平均温度和降水数据以及计算得到的月平均太阳辐射和潜在蒸散量为数据源,对甘肃省气候资源进行了区域化。对每种气象要素都采用了两种空间插值方法,并对插值结果运用了绝对验证和相对验证两种方法进行了验证和对比。结果表明:温度残差的平均绝对误差(MAE)是Spline< IDW, 其值分别为:0.744℃和0.754℃,平均相对误差(RME)分别为:9.56%和9.66%。降水的平均绝对误差是Kriging温度>潜在蒸散量>降水,但都达到了较高的精度。 相似文献
13.
人口是反映国情、国力基本情况的重要指标,是区域研究所必须考虑的重要因素之一。合理、准确地预测城市人口规模,是城市与区域规划中首先要考虑的基本问题,也是保证规划科学性与可实施性的关键性前提。以西宁市2000-2011年历年总人口为样本数据,分别构建了一元线性回归模型、马尔萨斯模型、logistic模型及GM（1,1）模型,并进行模型检验。结果表明：（1）模型均通过模型精度检验且精度较高,GM（1,1）模型拟合度最高,均误差达到0.004%,马尔萨斯模型拟合度最低,为-1.440 8%;（2）分析模型预测精度差异产生原因及适用性,表明深入、准确地分析样本数据特征,恰当选择分析方法对于控制人口预测精度尤为重要。由于西宁市2000-2011年人口样本数据在2005及2009年数据存在波动性,破坏了其与一元线性回归模型及马尔萨斯模型的拟合度,导致在4种模型中,Logistic及GM（1,1）模型预测精度较高,而GM（1,1）模预测精度最高,所以采用GM（1,1）模型进行西宁市人口预测,得到西宁市人口预测的最终结果：2012年西宁市总人口将达到225.89×104人,2015年将达到233.39×104人,2020年将达到246.37×104人。从结果看,未来9 a西宁市人口将呈现持续平稳增长的态势,但随着时间推进人口增长速度将逐渐下降。 相似文献
14.
Sand-dust weather has become an international social-environmental issue of common concern, and constitutes a serious threat to human lives and economic development. In order to explore the responses of natural desert sand and dust to the dynamics of water in desertification, we extracted long-term monitoring data related to precipitation, soil water, groundwater, and sand-dust weather. These data originated from the test stations for desertification control in desert areas of the middle reaches of the Heihe River. We used an algorithm of characteristic parameters, correlations, and multiple regression analysis to establish a regression model for the duration of sand-dust weather. The response characteristics of the natural desert sand and dust and changes of the water inter-annual and annual variance were also examined. Our results showed: (1) From 2006 to 2014 the frequency, duration, and volatility trends of sand-dust weather obviously increased, but the change amplitudes of precipitation, soil water, and groundwater level grew smaller. (2) In the vegetative growth seasons from March to November, the annual variance rates of the soil moisture content in each of four studied layers of soil samples were similar, and the changes in the frequency and duration of sand-dust weather were similar. (3) Our new regression equation for the duration of sand-dust weather passed the R test, F test, and t test. By this regression model we could predict the duration of sand-dust weather with an accuracy of 42.9%. This study can thus provide technological support and reference data for water resource management and research regarding sand-dust weather mechanisms. 相似文献
15.
GABRIELECRUCIANI MASSIMOBARONI SERGIOCLEMENTI GABRIELLECOSTANTINO DANIELARIGANELLI BERTSKAGERBERG 《地理学报(英文版)》1992,(6)
The standard deviation of prediction errors(SDEP)is used to evaluate and compare the predictive abilityof some regression models,namely MLR,ACE and linear and non-linear PLS,the last being the bestone.The parameter is determined by a cross-validation approach as an average of several runs obtainedon forming groups in a random way.The variation in SDEP with the number of latent variables in PLSis also discussed. 相似文献
16.
Yibin Ren Yang Zhang 《International journal of geographical information science》2019,33(9):1894-1912
Recently, researchers have introduced deep learning methods such as convolutional neural networks (CNN) to model spatio-temporal data and achieved better results than those with conventional methods. However, these CNN-based models employ a grid map to represent spatial data, which is unsuitable for road-network-based data. To address this problem, we propose a deep spatio-temporal residual neural network for road-network-based data modeling (DSTR-RNet). The proposed model constructs locally-connected neural network layers (LCNR) to model road network topology and integrates residual learning to model the spatio-temporal dependency. We test the DSTR-RNet by predicting the traffic flow of Didi cab service, in an 8-km2 region with 2,616 road segments in Chengdu, China. The results demonstrate that the DSTR-RNet maintains the spatial precision and topology of the road network as well as improves the prediction accuracy. We discuss the prediction errors and compare the prediction results to those of grid-based CNN models. We also explore the sensitivity of the model to its parameters; this will aid the application of this model to network-based data modeling. 相似文献
17.
土壤粒径（砂粒、粉粒和黏粒）是各种陆表过程和生态系统服务评估等模型的关键参数。作为一种土壤成分数据,土壤粒径的空间预测方法有和为1（或100%）等特殊要求,其空间分布精度受预测方法影响较大。本文针对土壤粒径相较于其他土壤属性的特殊性,提出了土壤粒径空间预测方法框架,综述了土壤粒径数据变换、空间插值和精度验证等系列方法,总结了提升土壤粒径空间预测精度的各种途径,包括通过有效的数据变换改善数据分布、结合数据分布特点选择合适的预测方法、结合辅助变量提升制图精度和分布合理性、使用混合模型提升插值精度、使用多成分联合模拟模型提升预测的系统性等。最后,提出了今后土壤粒径空间预测方法研究的未来方向,包括从考虑数据变换原理和机制角度改善数据分布、发展多成分联合模拟模型和高精度曲面建模方法,以及引入土壤粒径函数曲线并与随机模拟结合等。 相似文献
18.
Shraban Sarkar Archana K. Roy Tapas R. Martha 《International journal of geographical information science》2013,27(12):2436-2454
Soil formation depends upon several factors such as parent material, soil biota, topography and climate. It is difficult to use conventional soil survey methods for mapping the depth of soil in complex mountainous terrains. In this context, the present study aimed to estimate the soil depth for a large area (330.35 km2) using different geo-environmental factors through a soil-landscape regression kriging (RK) model in the Darjeeling Himalayas. RK with seven predictor variables such as elevation, slope, aspect, general curvature, topographic wetness index, distance from the streams and land use, was used to estimate the soil depth. While topographic parameters were derived from an 8-m resolution digital elevation model, the ortho-rectified Cartosat-1 satellite image was used to prepare the land use map. Soil depth measured at 148 sites within the study area was used to calibrate and validate the RK model. The result showed that the RK model with the seven predictors could explain 67% spatial variability of soil depth with a prediction variance between 0.23 and 0.42 m at the test site. In the regression analysis, land use (0.133) and slope (–0.016) were identified as significant determinants of soil depth. The prediction map showed higher soil depth in south-facing slopes and near valleys in comparison to other areas. Mean, mean absolute and root mean-square errors were used to access the reliability of the prediction, which indicated a goodness-of-fit of the RK model. 相似文献
19.
Yibin Ren Huanfa Chen Tao Cheng Yang Zhang Ge Chen 《International journal of geographical information science》2020,34(4):802-823
ABSTRACTThe spatio-temporal residual network (ST-ResNet) leverages the power of deep learning (DL) for predicting the volume of citywide spatio-temporal flows. However, this model, neglects the dynamic dependency of the input flows in the temporal dimension, which affects what spatio-temporal features may be captured in the result. This study introduces a long short-term memory (LSTM) neural network into the ST-ResNet to form a hybrid integrated-DL model to predict the volumes of citywide spatio-temporal flows (called HIDLST). The new model can dynamically learn the temporal dependency among flows via the feedback connection in the LSTM to improve accurate captures of spatio-temporal features in the flows. We test the HIDLST model by predicting the volumes of citywide taxi flows in Beijing, China. We tune the hyperparameters of the HIDLST model to optimize the prediction accuracy. A comparative study shows that the proposed model consistently outperforms ST-ResNet and several other typical DL-based models on prediction accuracy. Furthermore, we discuss the distribution of prediction errors and the contributions of the different spatio-temporal patterns. 相似文献
20.
Empirical models to predict the volumes of debris flows generated by recently burned basins in the western U.S. 总被引：3，自引：0，他引：3
Recently burned basins frequently produce debris flows in response to moderate-to-severe rainfall. Post-fire hazard assessments of debris flows are most useful when they predict the volume of material that may flow out of a burned basin. This study develops a set of empirically-based models that predict potential volumes of wildfire-related debris flows in different regions and geologic settings.The models were developed using data from 53 recently burned basins in Colorado, Utah and California. The volumes of debris flows in these basins were determined by either measuring the volume of material eroded from the channels, or by estimating the amount of material removed from debris retention basins. For each basin, independent variables thought to affect the volume of the debris flow were determined. These variables include measures of basin morphology, basin areas burned at different severities, soil material properties, rock type, and rainfall amounts and intensities for storms triggering debris flows. Using these data, multiple regression analyses were used to create separate predictive models for volumes of debris flows generated by burned basins in six separate regions or settings, including the western U.S., southern California, the Rocky Mountain region, and basins underlain by sedimentary, metamorphic and granitic rocks.An evaluation of these models indicated that the best model (the Western U.S. model) explains 83% of the variability in the volumes of the debris flows, and includes variables that describe the basin area with slopes greater than or equal to 30%, the basin area burned at moderate and high severity, and total storm rainfall. This model was independently validated by comparing volumes of debris flows reported in the literature, to volumes estimated using the model. Eighty-seven percent of the reported volumes were within two residual standard errors of the volumes predicted using the model. This model is an improvement over previous models in that it includes a measure of burn severity and an estimate of modeling errors. The application of this model, in conjunction with models for the probability of debris flows, will enable more complete and rapid assessments of debris flow hazards following wildfire. 相似文献