Research on Hydrological Time Series Prediction Based on Combined Model

Water level prediction of river runoff is an important part of hydrological forecasting.The change of water level not only has the trend and seasonal characteristics,but also contains the noise factors.And the water level prediction ability of a single model is limited.Since the traditional ARIMA(Autoregressive Integrated Moving Average)model is not accurate enough to predict nonlinear time series,and the WNN(Wavelet Neural Network)model requires a large training set,we proposed a new combined neural network prediction model which combines the WNN model with the ARIMA model on the basis of wavelet decomposition.The combined model fit the wavelet transform sequences whose frequency are high with the WNN,and the scale transform sequence which has low frequency is fitted by the ARIMA model,and then the prediction results of the above are reconstructed by wavelet transform.The daily average water level data of the Liuhe hydrological station in the Chu River Basin of Nanjing are used to forecast the average water level of one day ahead.The combined model is compared with other single models with MATLAB,and the experimental results show that the accuracy of the combined model is improved by 7%compared with the traditional wavelet network under the appropriate wavelet decomposition function and the combined model parameters.

NUMERICAL GEODYNAMIC MODELLING OF COUPLED MECHANO-THERMO-HYDROLOGICAL PROCESSES AND ITS APPLICATION IN PREDICTIVE EXPLORATION IN THE FENGHUANGSHAN ORE FIELD, TONGLING

Mineralisation is the result of the coupled multi-geodynamic processes in the crust. The coupled mechano-thermo-hydrological (MTH) processes are the basic physical processes that govern the location of the hydrothermal mineralization, which can be simulated in the computer by using of the numerical codes, such as FLAC. The numerical modeling results can be used not only to explain the features of existing ore deposits, but also to predict the favorable mineralization locations. This paper has summarized the basic equations describing coupled MHT processes in the water-saturated porous rocks, the principles of FLAC, and its application to the MHT processes related to copper mineralization in the Fenghuangshan ore field. We used the FLAC to simulate the syn-deformation cooling and fluid flowing evolution after the intrusion was emplaced and solidified. The modeling results suggest a most prospective exploration area where the subsequent exploration supported the prediction and the test bore hole disclosed the high quality copper ore bodies in the target, demonstrating a positive role of the numerical MTH modeling in facilitating predictive ore discovery.

Application of hydrological models in a snowmelt region of Aksu River Basin

This study simulated and predicted the runoff of the Aksu River Basin, a typical river basin supplied by snowmelt in an arid mountain region, with a limited data set and few hydrological and meteorological stations. Two hydrological models, the snowmelt-runoff model （SRM） and the Danish NedbФr-AfstrФmnings rainfall-runoff model （NAM）, were used to simulate daily discharge processes in the Aksu River Basin. This study used the snow-covered area from MODIS remote sensing data as the SRM input. With the help of ArcGIS software, this study successfully derived the digital drainage network and elevation zones of the basin from digital elevation data. The simulation results showed that the SRM based on MODIS data was more accurate than NAM. This demonstrates that the application of remote sensing data to hydrological snowmelt models is a feasible and effective approach to runoff simulation and prediction in arid unguaged basins where snowmelt is a major runoff factor.

Simulating and Prediction of Flow Using by WetSpa Model in Ziyarat River Basin, Iran

The spatially distributed hydrologic model WetSpa that works on daily, hourly, and minutely timescales is used to predict the flood hydrographs and spatial distribution of the hydrologic characteristics in a river basin by combining elevation, soil and land-use data within Geographical Information System. This model was applied in Ziarat river basin (95.15 km2) located in Golestan Province of Iran. Hourly hydro-meteorological data from 2008 to 2010 consist of precipitation data of two stations, temperature data of one station and evaporation data measured at one station, which were used as input data of the model. Three base maps namely DEM, land-use and soil types were produced in GIS form using 30 × 30 m cell size. Results of the simulations revealed a good agreement between calculated and measured hydrographs at the outlet of the river basin. The model predicted the hourly hydrographs with a good accuracy between 62% - 74% according to the Nash-Sutcliff criteria. To evaluate the model performance during the calibration and validation periods an Aggregated Measure (AM) was introduced that measures different aspects of the simulated hydrograph such as shape, size, and volume. The statistics of Ziarat river basin showed that the results produced by the model were very good in the calibration and validation periods.

基于TCN-BiLSTM与LSTM模型对比预测北洛河径流

【目的】本研究旨在探究TCN-BiLSTM耦合模型与传统LSTM模型在径流模拟预测中的性能,为洪水风险管理和区域水资源规划提供准确有效的径流预测模型。【方法】以北洛河流域为研究区,基于双向长短期记忆网络(BiLSTM)和时域卷积网络(TCN)建立一种新的径流预测耦合模型TCN-BiLSTM。利用相关性分析,筛选预测径流的输入因子,确定4种不同的输入方案应用于TCN-BiLSTM耦合模型和传统LSTM模型,每个模型分别预测1、2、3 d的径流量。采用平均绝对误差(MAE)、均方根误差(RMSE)和拟合优度(R^(2))来评估模型的预测性能。【结果】(1)TCN-BiLSTM耦合模型整体预测性能优于LSTM模型,TCN-BiLSTM模型R^(2)达到0.91,高于LSTM的0.89。相比于LSTM,TCN-BiLSTM对于峰值和突变点的捕捉能力更强,对于波动大的复杂数据预测效果更优;(2)在针对未来1~3 d径流量预测中,随着预见期的延长,4种方案下TCN-BiLSTM和LSTM模型的预测效果均有所下降,相较于预测1 d,预测3 d的TCNBiLSTM和LSTM模型的R^(2)分别平均下降了0.17和0.14,RMSE分别平均增大了4.59和4.40,MAE分别平均增大了1.26和1.31;(3)在4种输入方案里,日累积降水量和日径流量作为输入变量时,模型的预测效果最好。降水数据的加入使得TCN-BiLSTM和LSTM模型相较于单一日径流数据作为输入变量时,1、2、3 d径流量预测的R^(2)分别提高15%、14%、6%和18%、14%和1%。【结论】TCN-BiLSTM耦合模型和LSTM模型R^(2)均能达到0.85以上,TCN-BiLSTM模型R^(2)较LSTM提高了2%。对比来看,TCN-BiLSTM模型在拟合洪水过程中表现更为优异,对于汛期的预测性能优于非汛期。输入变量对模型的影响较大,有效且高质量的气象数据能够提高模型的预测性能。

基于SWAT模型的格尔木河上游分布式水文模拟和径流预测

研究格尔木河流域水文循环过程并预测未来流域水资源的变化特征,对地区生态环境保护和下游盐湖矿产资源可持续开发利用具有重要意义。选取格尔木水文站以上区域构建SWAT(Soil and Water Assessment Tool)分布式水文模型。采用大气同化数据集为气象驱动,联合区域内纳赤台和格尔木水文站的实测月尺度径流数据进行参数的率定和验证。在率定期和验证期内,纳什效率系数、确定性系数和相对偏差系数均达到了良好的标准,表明SWAT模型在格尔木河高寒山区流域水文过程模拟中具有较好的适用性。研究表明流域降水量偏少,地表径流量、壤中流量与降水量的变化趋势具有较好的一致性,降水量年际变化中蒸散发量为主要消耗量,占40.26%。根据未来气候预测模型RegCM4.6,预测路径浓度RCP2.6、RCP4.5和RCP8.53种情景下格尔木河未来40年径流量呈增加趋势。3种情景下的年平均径流量较基准期(2006—2018年)分别增加了7.63%、11.01%、15.96%;随着温室气体排放浓度的增加,径流量呈现出增加趋势,特别是夏秋季增幅较大。短时间内径流量增大可能会引发格尔木市洪涝灾害,破坏盐湖企业生产设施;但若将洪水资源进行调控和利用,不仅防范了洪涝灾害,同时也利于解决盐湖企业日渐增大的用水需求难题。

基于CMIP6模式的叶尔羌河流域未来水文干旱风险预估

全球变暖导致干旱等极端事件频发,严重威胁生态安全和社会经济可持续发展,尤其是干旱区这一气候响应敏感区。本研究基于流域站点观测气象、水文等数据和第六次国际耦合模式比较计划(CMIP6)全球气候模式数据,应用分布式水文模型HEC-HMS模型,模拟并预估了塔里木河重要源流——叶尔羌河流域历史(1986—2014年)与未来(2015—2100年)径流变化趋势以及水文干旱风险。结果表明:(1)HEC-HMS模型在干旱区流域适用性较好,未来3个共享社会经济路径(SSPs)下叶尔羌河流域年径流和标准化径流指数(SRI)都将呈显著增加趋势(P<0.1),SRI上升速率约0.13~0.27·(10a)^(-1)。(2)构建并对比了流域历史和未来时期4个干旱特征变量的边缘分布,未来干旱持续时间、干旱强度和烈度峰值都明显大于历史时期,干旱程度可能将不断加剧。(3)相较于历史时期,叶尔羌河流域未来水文干旱发生的联合概率将持续减小,且未来干旱事件的重现期都将有不同程度的延长。研究结果可为叶尔羌河流域水资源管理与规避水文干旱风险措施的制定提供科学参考。

极端气候事件对洞庭湖水文连通性变化的影响

为揭示洞庭湖近十几年水文连通性的变化特征并对其未来变化进行预测,研究极端气候事件在洞庭湖水文连通性变化中作出的贡献,使用水文连通性指数法、ETCCDI极端气候指数、Hurst指数以及数理统计分析方法进行研究。研究结果表明:洞庭湖水文连通性整体呈现夏季>秋季>春季>冬季的特征,夏季、秋季、春季和冬季的整体连通性指数均值分别为0.95、0.88、0.81和0.63,且洞庭湖的水文连通性在近30年比较稳定;经持续性预测发现洞庭湖水文连通性Hurst指数均大于0.5,表示其在没有人类活动干扰的情况下会在未来呈延续下降的趋势;洞庭湖水文连通性指数随着水位增加逐渐增加且增速逐渐放缓,水位增加对其具有正向影响的边际递减效应;极端降水事件对洞庭湖水文连通性具有较为明显的正向影响,而极端气温事件对洞庭湖水文连通性影响作用较小。研究结果有助于充分认识洞庭湖水资源演变规律,对保障洞庭湖流域水资源安全具有重要的理论和现实意义。

基于改进麻雀搜索算法的黄河花园口日径流预测研究

为准确预测河南省内黄河径流量,减少黄河泛滥隐患,帮助防洪工作开展。文章提出以改进麻雀搜索算法(ISSA)为基础的EMD-ISSA-LSTM径流预测模型。选取花园口水文站2009—2022年日径流数据作为实验数据,均方误差(MSE)、平均绝对百分比误差(MAPE)和纳什效率系数(NSE)作为模型评价指标。实验结果表明:EMDISSA-LSTM模型在花园口日径流预测中具有较好的准确性和稳定性,其中,预见期为1 d时NSE达到0.965。该研究为花园口水文站日径流预测工作提供了有效的工具,有利于水资源优化管理和水库的防洪调度。

基于LSTM的水文站流量短期预测建模差异性研究

当前水文预测模型研究缺乏对不同流域神经网络建模参数差异性选择的分析,模型的适应性较低,不利于模型推广运用。为此,选取黄河流域3个子流域为研究对象,基于离散小波算法和长短期记忆人工神经网络(LSTM)模型对不同特征子流域的水文站进行参数差异性建模研究,以提高水文预测模型的适应性。结果表明:历史流量周期性好、受人类活动影响小的水文站,可以基于水文站历史流量建立预测模型;对于受上游流量影响较大的干流水文站,预测模型仅依据水文站历史流量的预测性较低,若结合上游水文站流量,模型预测精度会有所提高,且考虑的上游水文站距预测水文站越远,模型可预测更长时间的流量;对于河道流量小、受地下水和降雨影响较大的流域,预测模型仅依据水文站历史流量的预测性较低,结合降水量可提高预测精度。

基于并行自适应遗传算法的水文模型率定研究

【目的】参数率定是影响水文模型预报精度的重要因素,采用人工智能算法可以有效提高水文模型参数的率定效果。【方法】采用基于种群离散程度的自适应算子,对GA算法的交叉、变异和迁移过程进行自适应优化,并利用粗粒度并行计算模型提高种群进化效率,综合以上手段研究了一种基于自适应策略的并行遗传算法。将传统遗传算法(GA),串行自适应遗传算法(AGA)和并行自适应遗传算法(PAGA),应用于屯溪流域新安江模型的参数率定,从率定效率、率定收敛性、率定稳定性和率定效果四个方面,验证PAGA算法的综合性能。【结果】结果表明:PAGA算法的计算加速效果显著,在10核环境下相对于AGA算法计算时间减少了87.9%;在进化后期,PAGA算法能够更加稳定的收敛于最优解,收敛后的目标函数值具有更好的稳定性;在验证期的场次洪水模拟中,采用PAGA算法率定的模型模拟效果最优,总体洪水合格率大于90%,确定性系数均值为0.85。【结论】PAGA算法能够明显降低模型参数寻优耗时,改善模型率定效果和收敛性能,为水文模型参数的率定提供了新思路。

Identification of Suitable Hydrologic Response Unit Thresholds for Soil and Water Assessment Tool Streamflow Modelling

Use of a non-zero hydrologic response unit(HRU) threshold is an effective way of reducing unmanageable HRU numbers and simplifying computational cost in the Soil and Water Assessment Tool(SWAT) hydrologic modelling. However, being less representative of watershed heterogeneity and increasing the level of model output uncertainty are inevitable when minor HRU combinations are disproportionately eliminated. This study examined 20 scenarios by running the model with various HRU threshold settings to understand the mechanism of HRU threshold effects on watershed representation as well as streamflow predictions and identify the appropriate HRU thresholds. Findings show that HRU numbers decrease sharply with increasing HRU thresholds. Among different HRU threshold scenarios, the composition of land-use, soil, and slope all contribute to notable variations which are directly related to the model input parameters and consequently affect the streamflow predictions. Results indicate that saturated hydraulic conductivity, average slope of the HRU, and curve number are the three key factors affecting stream discharge when changing the HRU thresholds. It is also found that HRU thresholds have little effect on monthly model performance, while evaluation statistics for daily discharges are more sensitive than monthly results. For daily streamflow predictions, thresholds of 5%/5%/5%(land-use/soil/slope) are the optimum HRU threshold level for the watershed to allow full consideration of model accuracy and efficiency in the present work. Besides, the results provide strategies for selecting appropriate HRU thresholds based on the modelling goal.

Seagull Lake,Western Eyre Peninsula,South Australia:A Saline Lake to Benefit from Climate Change? Ⅱ. Hydrology and Plants

Seagull Lake is an unusual saline lake,having a marine spring connected to a large continental ecosystem.With climate change the balance between marine,meteoric and groundwater inputs to,and evaporitic and groundwater

智能水文预报模型的研究与应用

洪水过程具有高度非线性、复杂性和非平稳性特征。将自适应步长的布谷鸟搜索(ASCS)算法应用于神经网络水文模型参数优化中,构建ASCS-LSTM洪水预报模型,并采用注意力机制进一步提高输入输出的相关性,实现高精度的智能洪水预测。在秦淮河流域的水位预测实验表明,ASCS-LSTM预报模型的预报结果要优于传统机器学习模型,稳定性和精确度得到提升,可为水文预报提供新思路。

水文气象学领域的水文循环研究进展

气候系统中的水循环处于不断运转演化和更新中。近年来在全球变暖和人类活动的双重影响下,水文循环发生了显著变化,引起了社会和学界的广泛关注。水文循环是气候系统的核心,是连接气候子系统的纽带,也是水文气象学研究的核心问题。近年来国内外学者采用各种观测手段及海-陆-气耦合模型检测和模拟水文循环变化,取得了丰硕成果。对其变化的物理机制和驱动因素有了深入的理解,提高了对其未来可能变化的预测预估水平。文中对最近20余年与水文气象学相关的水文循环发生的变化,引起全球、区域及流域水文循环通量变化的原因,以及未来变化的预测等问题所取得的进展做了较全面的阐述。最后对水文气象学领域水文循环变化研究应关注的问题进行了探讨。

基于指数核函数高斯过程回归的短期径流预测研究

径流预测有助于流域水资源综合高效调配和防洪减灾调度,如何精准地开展短期径流预测一直是水文水资源研究领域的重点。高斯过程回归(GPR)凭借其针对复杂非线性回归问题的泛化能力,已在水文过程长中短期预测研究中得到成功应用。而GPR回归分析能力不仅取决于模型参数,还受核函数影响。为此,研究分析了不同核函数作用下GPR预测模型效果,提出了基于指数核函数GPR的流域短期径流预测模型。首先通过多重相关性系数分析筛选相关性系数大且时段最短的预测因子组合,然后分别选用有理二次、径向基、马顿和指数核函数建立不同的GPR短期径流预测模型,同时加入了MLR、RT、SVM、BP等模型方法的预测结果作为对比。以赣江流域吉安水文站短期径流预测(预测步长为6 h,预见期为7 d)为例开展实例分析,相关实验结果表明:①应用不同核函数的GPR模型预测结果表现存在明显差异,不同方法预测表现由好到差分别为指数GPR、有理二次GPR、RT、马顿GPR、径向基GPR、SVM、MLR、BP;②指数GPR预测模型28时段的4项评价指标均表现最佳,DC和QR分别接近1和100%,预报精度达到甲级以上。综上,研究验证了指数核函数GPR短期径流预测模型的有效性和普适性,模型预测精度满足实际工程应用需求,具备实际应用价值。

水文变异对深度学习模型训练性能的影响研究

受气候变化和人类活动的双重影响,黄河流域水文情势发生剧烈演变,水文序列颠覆传统的“一致性”假设。为使深度学习模型更好适应于变异水文序列预测,采用Mann-Kendall检验方法确定径流时间序列变异点,对基于LSTM模型的k-fold交叉验证方法进行改进,提出考虑变异点的k-fold交叉验证方法,分析水文变异对深度学习模型训练的影响。选取黄河支流渭河华县站和黑河祁连山站为研究对象,对比分析不同变异程度的中长期水文序列预测结果。结果表明:渭河华县站的径流时间序列变异发生在1986年和2008年,黑河祁连山站径流时间序列变异发生在1988年;LSTM模型在中长期径流预测中,模型性能主要受时间序列长度的影响,其次受水文变化趋势及变异的影响。

面向产漂流性卵鱼类繁殖的水电站生态水文过程调控

基于水利工程影响下的产漂流性卵鱼类繁殖需求分析,开发了一种面向产漂流性卵鱼类繁殖的水电站生态水文过程调控方法,通过同步滚动调整启动时机和变动出库流量增幅来人工制造生态脉冲;在确定调控原则和目标的基础上,提出了水电站生态水文调控模型,并进行模型求解;在金沙江中游河段目标鱼类繁殖所需生态脉冲过程、中长期来水预测已知的前提下,将模型应用到梨园水电站生态水文调控实践,结果显示与模型类似工况的求解结果相近。实例验证了所提方法的可行性,可为其他已建水电站开展生态水文调控提供参考。

国内外流域水文模型研究进展

随着计算机的发展,流域水文模型种类越来越多。本文对流域水文模型进行分类,并整理细化出常见的水文模型,包括黑箱模型、概念性模型、物理模型以及集总式模型和分布式模型。其中,黑箱模型中比较典型的有总径流线性响应模型和人工神经网络,并在后者的基础上发展了BP-神经网络;概念性模型又包括萨克拉门托模型、水箱模型、新安江模型、SCS水文模型。针对不同类型水文模型,综述了对相关的原理、特点、优缺点、发展及应用的研究进展,为后续水文研究提供一定的依据和方向。

基于SWAT模型耦合精细化格点降水预报产品的径流预测研究

为探究复杂地形下水文预报方法,以贵州省平寨流域为例,采用阳长站2017~2019年逐日和2009~2018年逐月实测径流量数据,利用分布式SWAT水文模型进行建模、率定和验证,并耦合2021年6~9月流域气象关键点降水预报数据开展入库流量预报,最后采用水文模型评价指标分析了径流量预报效果。结果表明:SWAT模型在贵州省平寨流域有较好的适用性,能够较好地模拟实际径流量变化趋势与洪峰过程,率定期和验证期R 2和NSE系数均高于0.69;预报降水未订正时,预报入库流量在6~7月基本达标,但是8~9月效果不佳;预报降水订正后,预报入库流量在6~9月的NSE系数和R 2明显提高。