1.贵州省水利水电勘测设计研究院股份有限公司,贵州 贵阳 550002
2.河海大学水文水资源学院,江苏 南京 210098
李析男(1985—),男,博士,高级工程师,研究方向为水资源规划与管理。E-mail:lixinan1985@126.com
朱飞燕(2000—),女,硕士研究生,研究方向为水资源规划与管理、气候变化。E-mail:Zfyyyhhu@163.com
收稿:2025-06-03,
修回:2025-07-09,
录用:2025-07-09,
网络首发:2025-07-23,
纸质出版:2026-01-25
移动端阅览
李析男,朱飞燕.基于CNN-LSTM的贵州省水资源需水预测与趋势分析[J].人民珠江,2026,47(1):11-22.
LI Xi'nan,ZHU Feiyan.Forecast and Trend Analysis of Water Demand in Guizhou Province Based on CNN-LSTM[J].PEARL RIVER,2026,47(01):11-22.
李析男,朱飞燕.基于CNN-LSTM的贵州省水资源需水预测与趋势分析[J].人民珠江,2026,47(1):11-22. DOI: 10.3969/j.issn.1001-9235.2026.01.002.
LI Xi'nan,ZHU Feiyan.Forecast and Trend Analysis of Water Demand in Guizhou Province Based on CNN-LSTM[J].PEARL RIVER,2026,47(01):11-22. DOI: 10.3969/j.issn.1001-9235.2026.01.002.
水资源是影响经济发展、生态保护和社会稳定的关键因素,气候变化加剧了其不确定性,尤其在缺水或分布不均的地区。贵州省作为典型山区省份,地形复杂、气候多变,水资源管理面临挑战。基于2004—2023年贵州省水资源公报数据(降水量、水资源量、用水量) 和 CMIP6四个全球气候模式(BCC-CSM2-MR, CAMS-CSM1-0, CMCC-CM2-SR5, MIROC6)的未来气候数据,构建了 CNN-LSTM模型进行水资源供需预测,并分析了不同情景下水资源变化趋势。结果表明,CNN-LSTM模型预测精度高,测试集的平均绝对百分比误差为0.123 83(水资源量)和0.182 05(用水量),决定系数分别为0.990 63和0.990 67,表明模型能够有效捕捉数据的复杂时空变化趋势。未来预测显示,贵州省水资源量在 SSP245(中排放)和SSP585(高排放) 情景下均呈增加趋势,SSP585情景下变化幅度显著大于SSP245情景,且远期(2080—2099年)变化幅度远大于近期(2030—2049年),凸显了长期气候变化的累积效应。水资源量空间差异显著,黔南州(受地形抬升效应)增幅最大(远期SSP585达40亿m³),贵阳市(因高城镇化率削弱入渗)增幅最小(8.6亿m³)。需水量同样呈增长趋势,遵义市变化最明显(远期SSP585增加4.55亿m³)。气候变化对水资源的影响存在空间异质性,水资源分布不均和用水结构多样性是主要原因。模型预测未来水资源总量虽增加,但空间分布不均和用水结构差异带来的挑战依然严峻。 未来水资源管理应重视气候变化影响,推广节水技术,提高用水效率,建立预测模型,并针对不同区域特点(如黔南防洪、遵义需水管理)制定策略,确保水资源可持续利用。
Water resources are a key factor affecting economic development
ecological protection
and social stability. Climate change has exacerbated its uncertainty
especially in areas with water shortages or uneven distribution. As a typical mountainous province
Guizhou Province has complex terrain and a changeable climate
and water resources management is facing challenges. This article analyzed the data of
Guizhou Water Resources Bulletin
from 2004 to 2023 (precipitation
water resources amount
and water consumption) and future climate data of the four global climate models of CMIP6 (BCC-CSM2-MR
CAMS-CSM1-0
CMCC-CM2-SR5
and MIROC 6)
constructed the CNN-LSTM model to predict the supply and demand of water resources
and analyzed the changing trends of water resources under different scenarios. The results show that the CNN-LSTM model has high prediction accuracy. The average absolute percentage error (MAPE) of the test set is 0.123 83 (water resources) and 0.182
05 (water consumption)
and the decision coefficient (
R
2
) is 0.990 63 and 0.990 67 respectively
indicating that the model can effectively capture the trend of complex space-time changes of data. Future forecasts show that the amount of water resources in Guizhou Province is increasing under both SSP245 (medium emissions) and SSP585 (high emissions) scenarios
and the change under the SSP585 scenario is significantly greater than that of the SSP245 scenario; the change in the far future (2080–2099) is much greater than that in the near future (2030–2049)
highlighting the cumulative effect of long-term climate change. The spatial difference in the amount of water resources is significant. Qiannan Prefecture (due to the topographic elevation effect) has the largest increase (long-term SSP585 reaches 4 billion m
3
)
and Guiyang City (weak penetration due to the high urbanization rate) has the smallest increase (860 million m
3
). Water demand is also on the rise
and the change in Zunyi City is the most obvious (long-term SSP585 has increased by 455 million m
3
). The impact of climate change on water resources is spatially heterogeneous. The uneven distribution of water resources and the diversity of water structures are the main reasons. The model predicts that although the total amount of water resources will increase in the future
the challenges posed by uneven spatial distribution and differences in water structure are still serious. In the future
water resources management should pay attention to the impact of climate change
promote water conservation technologies
improve water efficiency
establish predictive models
and formulate strategies for different regional characteristics (such as flood control in southern Guizhou and water demand management in Zunyi) to ensure the sustainable use of water resources.
傅懿兵 . 浅谈水资源科学管理对促进经济可持续发展的重要性 [J]. 给水排水 , 2022 , 58 ( 8 ): 212 - 213 .
FU Y B . A Brief Discussion on the Importance of Scientific Water Resource Management in Promoting Sustainable Economic Development [J]. Water & Wastewater Engineering , 2022 , 58 ( 8 ): 212 - 213 . (in Chinese)
LI J L , TIAN P , SHI X , et al . Research progress on water resource management [J]. Journal of Zhejiang University(Science Edition) , 2019 , 46 : 248 - 260 .
CIAMPITTIELLO M , MARCHETTO A , BOGGERO A . Water Resources Management under Climate Change: A Review [J]. Sustainability , 2024 , 16 : 35 - 39 .
KALFAS D , KALOGIANNIDIS S , PAPAEVANGELOU O , et al . Assessing the Connection between Land Use Planning, Water Resources, and Global Climate Change [J]. Water , 2024 , 16 ( 2 ): 333 - 343 .
IPCC第六次评估报告第一工作组报告发布 [N]. 中国气象报 , 2021-08-10 (1). doi: 10.1017/9781009325844.031 http://dx.doi.org/10.1017/9781009325844.031
IPCC Sixth Assessment Report Working Group I Report Released [N]. China Meteorological News , 2021-08-10 (1). (in Chinese)
龙晓彤 , 曹艳青 . 黄河流域水资源利用时空演变特征及驱动要素 , 2024 : 355 - 362 .
LONG X T , CAO Y Q . Spatiotemporal Evolution Characteristics and Driving Factors of Water Resources Utilization in the Yellow River Basin , 2024 (12th) China Water Ecology Conference Proceedings , 2024 : 355 - 362 . (in Chinese)
虞畅 , 金君良 , 王国庆 , 等 . 黄河源区典型流域水文要素演变及其对气候变化的响应 [J]. 南水北调与水利科技(中英文) , 2024 , 22 ( 5 ): 845 - 855 .
YU C , JIN J L , WANG G Q , et al . Evolution of Hydrological Elements in Typical Basins of the Yellow River Source Region and Their Response to Climate Change [J]. South-to-North Water Transfers and Water Science & Technology (Chinese and English) , 2024 , 22 ( 5 ): 845 - 855 . (in Chinese)
王国庆 , 虞畅 , 金君良 , 等 . 气候变化对南水北调西线工程的影响及建议 [J]. 人民黄河 , 2024 , 46 ( 9 ): 60 - 66 .
WANG G Q , YU C , JIN J L , et al . Impacts of Climate Change on the Western Route of the South-to-North Water Diversion Project and Recommendations [J]. Yellow River , 2024 , 46 ( 9 ): 60 - 66 . (in Chinese)
李析男 , 何淋虹 , 薛联青 . 气候变化下贵州省极端气候事件与植被覆盖的时空分异及预测 [J]. 人民珠江 , 2024 , 45 ( 10 ): 65 - 75 .
LI X N , HE L H , XUE L Q . Spatiotemporal Differentiation and Prediction of Extreme Climate Events and Vegetation Coverage in Guizhou Province under Climate Change [J]. Pearl River , 2024 , 45 ( 10 ): 65 - 75 . (in Chinese)
龚骁 , 戴长雷 , 李树岭 , 等 . 基于CMIP6气候情景的松花江流域典型支流水文模拟及预测:以蚂蚁河流域为例 [J]. 人民珠江 , 2025 , 46 ( 11 ): 65 - 74 .
GONG X , DAI C L , LI S L , et al . Hydrological Simulation and Prediction of Typical Tributaries in the Songhua River Basin Based on CMIP6 Climate Scenarios—A Case Study of the Mayi River Basin [J]. Pearl River , 2025 , 46 ( 11 ): 65 - 74 . (in Chinese)
PANDA K C , SINGH R M , THAKURAL L N , et al . Representative grid location-multivariate adaptive regression spline (RGL-MARS) algorithm for down scaling dry and wet season rainfall [J]. Journal of hydrology , 2022 , 605 . DOI: 10.1016/j.jhydrol. 2021. 127381 http://dx.doi.org/10.1016/j.jhydrol.2021.127381 .
ALMAZROUI M , SAEED F , SAEED S , et al . Projected change in temperature and precipitation over Africafrom CMIP6 [J]. Earth Systems and Environment , 2020 , 4 : 455 - 475 .
SOBIE S R , ZWIERS F W , CURRY C L . Climate model projections for Canada: a comparison of CMIP5 and CMIP6 [J]. Atmosphere-Ocean , 2021 , 59 ( 4/5 ): 269 - 284 .
LECUN Y , BOTTOU L , BENGIO Y , et al . Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE , 2002 , 86 ( 11 ): 2278 - 2324 .
HE K M , ZHANG X Y , REN S Q , et al . Identity mappings in deep residual networks [C]// Computer Vision-ECCV 2016: 14th European Conference , Amsterdam, The Netherlands. Springer International Publishing , 2016 : 630 - 645 .
SHI X J , CHEN Z R , WANG H , et al . Convolutional LSTM network: A machine learning approach for precipitation nowcasting [J]. Advances in neural information processing systems , 2015 , 28 ( 5 ): 112 - 132 .
PU Z H , YAN J R , CHEN L , et al . A hybrid Wavelet-CNN-LSTM deep learning model for short-term urban water demand forecasting [J]. Frontiers of Environmental Science & Engineering , 2023 , 17 ( 2 ): 22 - 49 .
李轩 , 吴永强 , 王佳伟 , 等 . 基于SVM-SARIMA-LSTM模型的城市用水量实时预测 [J]. 水电能源科学 , 2025 , 43 ( 3 ): 36 - 39 .
LI X , WU Y Q , WANG J W , et al . Real-time urban water consumption prediction based on SVM-SARIMA-LSTM model [J]. Water Resources and Power , 2025 , 43 ( 3 ): 36 - 39 . (in Chinese)
陈星 , 沈紫菡 , 许钦 , 等 . 基于CNN-LSTM-Attention的月生活需水预测研究 [J]. 三峡大学学报(自然科学版) , 2024 , 46 ( 5 ): 1 - 6 .
CHEN X , SHEN Z H , XU Q , et al . Research on Monthly Domestic Water Demand Prediction Based on CNN-LSTM-Attention [J]. Journal of China Three Gorges University (Natural Science Edition) , 2024 , 46 ( 5 ): 1 - 6 . (in Chinese)
WEN H , MA H C , DU Y , et al . Prediction of irrigation water requirement based on parallel CNN-LSTM model and Mann-Kendall test [C]// 2024 IEEE 7th International Conference on Industrial Cyber-Physical Systems (ICPS) . IEEE , 2024 : 1 - 6 .
陈金红 , 崔东文 . 基于深度学习神经网络超参数优化的入库径流预测方法研究:以云南省暮底河水库为例 [J]. 三峡大学学报(自然科学版) , 2023 , 45 ( 4 ): 25 - 32 .
CHEN J H , CUI D W . Research on Inflow Runoff Prediction Method Based on Deep Learning Neural Network Hyperparameter Optimization—A Case Study of Mudi River Reservoir in Yunnan Province [J]. Journal of China Three Gorges University (Natural Science Edition) , 2023 , 45 ( 4 ): 25 - 32 . (in Chinese)
HOCHREITER S , SCHMIDHUBER J . Long short-term memory [J]. Neural Computation , 1997 , 9 ( 8 ): 1773 - 1780 .
LI H B , SHEFFIELD J , WOOD E F . Bias correction of monthly precipitation and temperature fields from Intergovernmental Panel on Climate Change AR4 models using equidistant quantile matching [J]. Journal of Geophysical Research: Atmospheres , 2010 , 15 ( 10 ): 23 - 32 .
SU F , ZHANG L , OU T , et al . Hydrological response to future climate changes for the major upstream river basins in the Tibetan Plateau [J]. Global and Planetary Change , 2016 , 136 : 82 - 95 .
周嘉月 , 卢麾 , 阳坤 , 等 . 基于CMIP6的中高温升情景对中国未来径流的预估 [J]. 中国科学:地球科学 , 2023 , 53 ( 3 ): 505 - 524 . doi: 10.1007/s11430-022-1055-5 http://dx.doi.org/10.1007/s11430-022-1055-5
ZHOU J Y , LU H , YANG K , et al . Projection of future runoff in China under medium-to-high warming scenarios based on CMIP6 [J]. Science China Earth Sciences , 2023 , 53 ( 3 ): 505 - 524 . (in Chinese)
张佳怡 , 伦玉蕊 , 刘浏 , 等 . CMIP6多模式在青藏高原的适应性评估及未来气候变化预估 [J]. 北京师范大学学报(自然科学版) , 2022 , 58 ( 1 ): 77 - 89 .
ZHANG J Y , LUN Y R , LIU L , et al . CMIP6 Multi-Model Evaluation and Future Climate Change Projections over the Tibetan Plateau [J]. Journal of Beijing Normal University (Natural Science Edition) , 2022 , 58 ( 1 ): 77 - 89 . (in Chinese)
O'NEILL B C , TEBALDI C , VAN VUUREN D P , et al . The scenario model intercomparison project (ScenarioMIP) for CMIP6 [J]. Geoscientific Model Development , 2016 , 9 ( 9 ): 3461 - 3482 .
0
浏览量
17
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621