基于SSA-BP神经网络模型的全球入海径流量未来变化趋势

doi:10.12006/j.issn.1673-1719.2023.243

摘要/Abstract

摘要：

入海径流是水循环的重要环节，探究气候变化背景下全球入海径流量的时空变化特征，可为水资源合理利用提供依据。基于全球376条外流河逐月流量、ERA5-LAND再分析资料和10个全球气候模式，构建基于SSA-BP神经网络的降水径流关系模型，分析全球入海径流量在历史时期（1961—2020年）和未来（2021—2100年）3种情景（SSP1-2.6、SSP3-7.0和SSP5-8.5）下的时空变化特征。研究发现：(1)全球尺度上，1961—2020年，多年平均入海年径流量为37423 km³。2021—2100年，全球入海年径流量在未来3种情景下均呈增加趋势，SSP1-2.6情景下趋势显著。与基准期相比，21世纪末期增幅最大。(2)洲际尺度上，历史时期，非洲入海径流量呈显著减少趋势，北美洲呈显著增加趋势。2021—2100年，亚洲、北美洲在3种情景下呈增加趋势，大洋洲呈减少趋势，其余各大洲情景间差异明显。(3)纬向分布上，历史时期，南北半球低纬度变化趋势不显著；北半球中纬度呈弱减少趋势，南半球中纬度呈显著减少趋势；北半球高纬度呈显著增加趋势。2021—2100年，从低到高排放情景，入海径流在北半球低纬度的增加趋势和在南半球低纬度的减少趋势愈发显著；北半球中高纬由低排放情景的显著增加转变为中高排放情景的显著减少；南半球中纬度在低排放情景下呈显著增加趋势，在中高排放情景下趋势不显著。

关键词: 入海径流量, 趋势预估, SSA-BP模型, 全球

Abstract:

Runoff into the sea is an important link in the water cycle. Exploring the spatial and temporal variation characteristics of global runoff into the sea under the background of climate change can provide a basis for the rational utilization of water resources. Based on the monthly discharge of 376 outflow rivers around the world, ERA5-LAND reanalysis data and 10 global climate models, a precipitation-runoff relationship model based on the SSA-BP neural network was constructed to analyze the spatiotemporal change characteristics of the global runoff into the sea during the historical period (1961-2020) and the future (2021-2100) under three scenarios (SSP1-2.6, SSP3-7.0 and SSP5-8.5). The results are as follows. (1) On a global scale, from 1961 to 2020, the multi-year average annual runoff into the sea was 37423 km³. From 2021 to 2100, the global annual runoff into the sea will show an increasing trend under the three future scenarios, with a significant trend under the SSP1-2.6 scenario. Compared with the base period, the late 21st century showed the largest increase. (2) On an intercontinental scale, during historical periods, Africa’s runoff into the sea showed a significant decreasing trend, while North America showed a significant increasing trend. From 2021 to 2100, Asia and North America will show an increasing trend under three scenarios, Oceania will show a decreasing trend, and there are obvious differences among the scenarios for the remaining continents. (3) In terms of latitudinal distribution, the change trend in the low latitudes of the Northern and Southern Hemispheres during the historical period was not significant; the mid-latitude Northern Hemisphere showed a weak decreasing trend, the Southern Hemisphere showed a significant decreasing trend; and the high latitudes of the Northern Hemisphere showed a significant increasing trend. From 2021 to 2100, from low to high emission scenarios, the increasing trend of runoff into the sea in the low latitudes of the Northern Hemisphere and the decreasing trend in the Southern Hemisphere become more and more significant; in the mid-to-high latitudes of the Northern Hemisphere, the significant increase in the low emission scenario changes to the significant decrease in the medium and high emission scenario; The mid-latitudes of the Southern Hemisphere show a significant increasing trend under the low emission scenario, but the trend is not significant under the medium and high emission scenario.

Key words: Runoff into the sea, Trend estimation, SSA-BP model, Global scale

赵鹏, 姜彤, 苏布达, 高妙妮. 基于SSA-BP神经网络模型的全球入海径流量未来变化趋势[J]. 气候变化研究进展, 2024, 20(2): 182-192.

ZHAO Peng, JIANG Tong, SU Bu-Da, GAO Miao-Ni. Research on the future change trend of global runoff into the sea based on SSA-BP neural network model[J]. Climate Change Research, 2024, 20(2): 182-192.

图/表 9

图1 控制性水文站

Fig. 1 Controlled hydrological stations

图2 BP神经网络结构(a)和SSA-BP神经网络模型计算流程(b)

Fig. 2 BP neural network structure (a) and SSA-BP neural network model calculation process (b)

表1 SSA-BP神经网络模型模拟与验证

Table 1 SSA-BP neural network model simulation and verification

图3 1961—2020年全球入海径流量年际(a)、年代际(b)、月(c)和季节(d)变化注：(b)图距平值是相对于1961—2020年多年平均计算得出。

Fig. 3 Inter-annual (a), inter-decadal (b), monthly (c) and seasonal (d) changes in global runoff into the sea from 1961 to 2020

图4 2021—2100年全球入海年径流量(a)和各月径流量(b)变化注：实线表示多模式集合，阴影表示多模式的上下限。

Fig. 4 Changes of annual (a) and monthly (b) global runoff into the sea from 2021 to 2100. (The solid line represents the multi-model ensemble, and the shaded represents the upper and lower bounds of the multiple models)

图5 历史时期和未来不同情景下全球主要外流河变化态势

Fig. 5 The changes in major global outflow rivers during historical period and under different scenarios in the future

图6 1961—2020年全球各大洲各月入海径流量变化

Fig. 6 The changes of monthly runoff into the sea from all continents around the world from 1961 to 2020

表2 2021—2100年不同情景下全球各大洲入海径流量年际变化

Table 2 Inter-annual changes in runoff into the sea from all continents in the world under different scenarios from 2021 to 2100 km3/(10 a)

表3 2021—2100年不同情景下全球各纬度带入海径流量年际变化

Table 3 Inter-annual changes in runoff into the sea at various latitudes around the world under different scenarios from 2021 to 2100 km3/(10 a)

参考文献 29

[1]	IPCC. Climate change 2022: impacts, adaptation and vulnerability[M]. Cambridge: Cambridge University Press, 2022
[2]	Nijssen B, O′Donnell G M, Lettenmaier D P, et al. Predicting the discharge of global rivers[J]. Journal of Climate, 2001, 14 (15): 3307-3323 doi: 10.1175/1520-0442(2001)014<3307:PTDOGR>2.0.CO;2 URL
[3]	Fekete B M, Vörösmarty C J, Grabs W. High-resolution fields of global runoff combining observed river discharge and simulated water balances[J]. Global Biogeochemical Cycles, 2002, 16 (3): 1-10 doi: 10.1029/2001GB001398 URL
[4]	Alcamo J, Döll P, Henrichs T, et al. Development and testing of the WaterGAP 2 global model of water use and availability[J]. Hydrological Sciences Journal, 2003, 48 (3): 317-337 doi: 10.1623/hysj.48.3.317.45290 URL
[5]	Dai A, Trenberth K E. Estimates of freshwater discharge from continents: latitudinal and seasonal variations[J]. Journal of Hydrometeorology, 2002, 3 (6): 660-687 doi: 10.1175/1525-7541(2002)003<0660:EOFDFC>2.0.CO;2 URL
[6]	Syed T H, Famiglietti J S, Chambers D P, et al. DGRACE-based estimates of terrestrial freshwater discharge from basin to continental scales[J]. Journal of Hydrometeorology, 2009, 10 (1): 22-40 doi: 10.1175/2008JHM993.1 URL
[7]	Clark E, Sheffield J, Nijssen E, et al. Continental runoff into the oceans (1950-2008)[J]. Journal of Hydrometeorology, 2015, 16 (4): 1502-1520 doi: 10.1175/JHM-D-14-0183.1 URL
[8]	梁书民, 于智媛. 用经验径流系数推算全球径流深度分布场[J]. 干旱区研究, 2018, 35 (1): 1-11.
	Liang S M, Yu Z Y. Global runoff field calculated by empirical runoff coefficient[J]. Arid Zone Research, 2018, 35 (1): 1-11 (in Chinese)
[9]	Oki T, Kanae S. Global hydrological cycles and world water resources[J]. Science, 2006, 313 (5790): 1068-1072 doi: 10.1126/science.1128845 pmid: 16931749
[10]	IPCC. Climate change 2021: the physical science basis[M]. Cambridge: Cambridge University Press, 2021
[11]	Liang Y C, Lo M H, Lan C W, et al. Amplified seasonal cycle in hydroclimate over the Amazon River basin and its plume region[J]. Nature Communications, 2020, 11 (1): 4390 doi: 10.1038/s41467-020-18187-0
[12]	秦恺, 张瑜, 高郭平, 等. 北冰洋入海径流变化状况及其与近岸海冰关系[J]. 极地研究, 2019, 31 (4): 371-382. doi: 10.13679/j.jdyj.20180055
	Qin K, Zhang Y, Gao G P, et al. Variation of Arctic ocean runoff and its correlation with coastal sea ice[J]. Chinese Journal of Polar Research, 2019, 31 (4): 371-382 (in Chinese) doi: 10.13679/j.jdyj.20180055
[13]	O′Neill B C, Tebaldi C, Vuuren D P, et al. The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6[J]. Geoscientific Model Development, 2016, 9 (9): 3461-3482 doi: 10.5194/gmd-9-3461-2016 URL
[14]	Zhang Y, Zheng H X, Zhang X Z, et al. Future global streamflow declines are probably more severe than previously estimated[J]. Nature Water, 2023, 1 (3): 261-271 doi: 10.1038/s44221-023-00030-7
[15]	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.
	Zhou J Y, Lu H, Yang K, et al. Projection of China’s future runoff based on the CMIP6 mid-high warming scenarios[J]. Science China: Earth Sciences, 2023, 53 (3): 505-524 (in Chinese)
[16]	韩振宇, 徐影, 吴佳, 等. 多区域气候模式集合对中国径流深的模拟评估和未来变化预估[J]. 气候变化研究进展, 2022, 18 (3): 305-318.
	Han Z Y, Xu Y, Wu J, et al. Evaluation on the simulated runoff in China and future change projection based on multiple regional climate models[J]. Climate Change Research, 2022, 18 (3): 305-318 (in Chinese)
[17]	Zeng P. Artificial neural networks principle for finite element method[J]. Journal of Applied Mathematics and Mechanics, 1996, 76 (1996): 565-566
[18]	Rumelhart D, Hinton G E, Williams R J. Learning representations by Back-Propagating errors[J]. Nature, 1986, 323 (1986): 533-536 doi: 10.1038/323533a0
[19]	李志新, 赖志琴. 年径流变化的BP神经网络预报模型研究[J]. 水电能源科学, 2018, 36 (7): 10-12.
	Li Z X, Lai Z Q. Study on BP neural network forecast model for annual runoff change[J]. Water Resources and Power, 2018, 36 (7): 10-12 (in Chinese)
[20]	农振学, 王超, 雷晓辉. 基于主成分分析和BP神经网络的赣江流域中长期径流预报[J]. 水电能源科学, 2018, 36 (1): 16-19.
	Nong Z X, Wang C, Lei X H. Long-term runoff forecasting in Ganjiang River basin based on principal component analysis and BP neural network[J]. Water Resources and Power, 2018, 36 (1): 16-19 (in Chinese)
[21]	陈雪怡, 陈元芳, 王文鹏, 等. 月径流预报建模方法对比分析: 以嘉陵江北碚站为例[J]. 人民长江, 2022, 53 (9): 80-86.
	Chen X Y, Chen Y F, Wang W P, et al. Evolution features of drought: flood in Henan province and their climatic background during 1960 to 2018[J]. Yangtze River, 2022, 53 (9): 80-86 (in Chinese)
[22]	Xue J K, Shen B. A novel swarm intelligence optimization approach: sparrow search algorithm[J]. Systems Science & Control Engineering, 2020, 8 (1): 22-34
[23]	葛渊博, 卢文喜, 白玉堃, 等. 基于SSA-BP与SSA的地下水污染源反演识别[J]. 中国环境科学, 2022, 42 (11): 5179-5187.
	Ge Y B, Lu W X, Bai Y K, et al. Inversion and identification of groundwater pollution sources based on SSA-BP and SSA[J]. China Environmental Science, 2022, 42 (11): 5179-5187 (in Chinese)
[24]	Global Runoff Data Center (GRDC). Major river basins of the world Global Runoff Data Centre[R]. Germany: Federal Institute of Hydrology (BfG), 2020: 1-7
[25]	Bain R L, Shaw M J, Geheran M P, et al. Intercomparison of global ERA reanalysis products for streamflow simulations at the high-resolution continental scale[J]. Journal of Hydrology, 2023, 616: 128624 doi: 10.1016/j.jhydrol.2022.128624 URL
[26]	Joaquín S M, Emanuel D, Anna P A, et al. ERA5-Land: a state-of-the-art global reanalysis dataset for land applications[J]. Earth System Science Data, 2021, 13 (9): 4349-4383 doi: 10.5194/essd-13-4349-2021 URL
[27]	曹丽格, 方玉, 姜彤, 等. IPCC影响评估中的社会经济新情景(SSPs)进展[J]. 气候变化研究进展, 2012, 8 (1): 74-78.
	Cao L G, Fang Y, Jiang T, et al. Advances in shared socio-economic pathways for climate change research and assessment[J]. Climate Change Research, 2012, 8 (1): 74-78 (in Chinese)
[28]	张丽霞, 陈晓龙, 辛晓歌. CMIP6 情景模式比较计划(ScenarioMIP)概况与评述[J]. 气候变化研究进展, 2019, 15 (5): 519-525.
	Zhang L X, Chen X L, Xin X G. Short commentary on CMIP6 Scenario Model Intercomparison Project (ScenarioMIP)[J]. Climate Change Research, 2019, 15 (5): 519-525 (in Chinese)
[29]	魏凤英. 现代气候统计诊断与预测技术(第2版)[M]. 北京: 气象出版社, 2007: 37-105.
	Wei F Y. Climatological statistical diagnosis and prediction technology (second edition)[M]. Beijing: China Meteorological Press, 2007: 37-105 (in Chinese)