Applicability of the random forest model in quantifying the attribution of runoff changes
-
摘要: 气候变化和人类活动是影响流域径流变化的两大因素, 定量识别二者对流域径流的影响对水资源合理开发和流域综合治理具有重要意义。随机森林模型作为一种易于使用的机器学习方法, 被越来越多地应用于水文学领域, 然而随机森林模型在径流变化归因分析中是否具有适用性值得探讨。本文以永定河上游的洋河流域为例, 基于随机森林模型对流域径流进行了模拟, 采用径流影响评价模型定量分析了气候变化和人类活动对径流变化的贡献率; 并利用基于物理机制的SWAT模型对研究结果进行了对比验证。研究结果表明: 1)在对不同子流域径流模拟效果方面, SWAT分布式水文模型对东洋河(柴沟堡东站)、南洋河(柴沟堡南站)、洋河(响水堡站)流域突变前的径流量模拟结果较为可信, 3个流域率定期和验证期的R2均在0.65以上, 纳什系数(NSE)也大部分在0.65以上; 随机森林模型对3个流域模拟径流的NSE和R2多在0.80以上, 均高于SWAT模型的NSE和R2, 随机森林模型模拟表现优于SWAT模型; 2)对比验证发现, 基于随机森林模型和SWAT模型的流域径流变化归因分析结果较为相近, 人类活动是导致永定河上游流域径流变化的主要原因, 对径流减少的贡献率为84.3%~97.6%。总体上, 随机森林模型在永定河上游流域径流变化归因研究中具有一定的适用性, 为流域径流变化贡献率的定量识别提供了一种新的方法和思路。Abstract: Climate change and human activity have a significant impact on runoff in basins. As an important ecological barrier,the upper Yongding River Basin of has undergone significant changes in the ecological environment over the past 50 years, and the problem of water shortage has become increasingly prominent. It is necessary to restore the water ecology and analyze the influence of climate and human activities on runoff dynamics. Therefore, this study established a comparative approach between the random forest and Soil and Water Assessment Tool (SWAT) models in the Yanghe River Basin, which is greatly affected by human activities in the upper Yongding River Basin. The main conclusions were as follows: 1) in terms of the runoff simulation effect, the SWAT model was reliable for revealing the runoff dynamics in the Dongyanghe River Basin, Nanyanghe River Basin, and Yanghe River Basin. The R2 values of the simulated and observed runoff in the three basins were above 0.65 in both the calibration and verification periods, and the Nash coefficients (NSE) were also above 0.65 in the three basins. However, the random forest model outperformed the SWAT model in terms of NSE and R2 in the three basins, and its NSE and R2 values were mostly above 0.80. 2) In quantifying the attribution of runoff changes, the results based on the SWAT model showed that the contribution rates of climate change to runoff decline in the three basins were generally between 5.0% and 15.7%, and those of human activities were 84.3%–95.0% in the three basins. The results based on the random forest model were similar to the attribution results of runoff decline based on the SWAT model; the contribution rates of climate change and human activities to runoff decline in the three basins were generally 2.4%–11.5% and 88.5%–97.6% in the three basins. This is consistent with the research results of other experts and scholars that human activities are the main cause of runoff decline in the Yanghe River Basin. Random forest can be applied in runoff simulation in the Yanghe River Basin, and the simulated model results can be used in water resource management. In this study, the SWAT model and random forest were combined to reveal the impacts of climate change and human activities on the changes in runoff in the Yanghe River Basin. Additionally, the applicability of the random forest model in the Yanghe River Basin was evaluated, which demonstrated the possibility of integrating the random forest model in hydrological modeling in further research. However, random forest is a black-box model in theory and lacks consideration of hydrological processes. Although this study preliminarily explored the method and it has been proven to be applicable in runoff simulation, the uncertainty of this method in runoff simulation or runoff evolution needs to be further explored.
-
Key words:
- Yongding River Basin /
- Random forest model /
- SWAT /
- Runoff /
- Attribution analysis
-
图 2 基于SWAT模型的东洋河、南洋河、洋河流域突变前月径流实测值和模拟值对比
图中虚线左侧为率定期, 右侧为验证期。
Figure 2. Observed and simulated monthly runoff before abrupt change in the Dongyanghe River Basin, Nanyanghe River Basin, Yanghe River Basin based on SWAT
In the figure, the left side of the dotted line is the calibration period, and the righ side of the dotted line is the verification period.
表 1 洋河流域水文站、气象站基本信息
Table 1. Basic information of hydrological stations and weather stations in the Yanghe River Basin
分类 Category 站名 Name 经度 Longitude (°) 纬度 Latitude (°) 海拔/控制面积 Altitude (m) /area (km²) 时间序列 Period 气象站
Meteorological station天镇 Tianzhen 114.1 40.4 1015.6 1984—2015 尚义 Shangyi 114.0 41.1 1358.6 1984—2015 怀安 Huai’an 114.4 40.7 1142.4 1984—2015 宣化 Xuanhua 115.0 41.6 629.3 1984—2015 崇礼 Chongli 115.3 41.0 1246.7 1984—2015 涿鹿 Zhuolu 115.2 40.3 529.6 1984—2015 兴和 Xinghe 113.9 40.9 1253.1 1984—2015 水文站
Hydrometric station柴沟堡东 Chaigoubu East 114.1 40.7 3674 1984—2015 柴沟堡南 Chaigoubu South 114.4 40.7 2427 1984—2015 响水堡 Xiangshuibu 115.2 40.5 14 507 1984—2015 表 2 基于SWAT模型的气候变化及人类活动对东洋河、南洋河、洋河流域径流变化影响的贡献率(1986—2015年)
Table 2. Contribution rates of climate change and human activities to runoff changes in the Dongyanghe River Basin, Nanyanghe River Basin, Yanghe River Basin based on SWAT model (1986 to 2015)
水文站
Hydrometric station时段
Time interval实测径流
Observed
runoff
(×108 m3)模拟径流
Simulated
runoff
(×108 m3)径流变化量
Runoff
variation
(×108 m3)贡献值
Variable quantity (×108 m3)贡献率
Contribution rate (%)气候变化
Climate change人类活动
Human activities气候变化
Climate change人类活动
Human activities柴沟堡东
Chaigoubu East Station突变前
Before abrupt change0.93 突变后
After abrupt change0.42 0.85 −0.51 −0.08 −0.43 15.7 84.3 柴沟堡南
Chaigoubu South Station突变前
Before abrupt change0.47 突变后
After abrupt change0.19 0.45 −0.28 −0.01 −0.27 5.0 95.0 响水堡
Xiangshuibu Station突变前
Before abrupt change2.14 突变后
After abrupt change0.95 2.04 −1.20 −0.11 −1.09 8.9 91.1 表 3 基于随机森林模型的气候变化和人类活动对东洋河、南洋河、洋河流域径流变化的贡献率(1986—2015年)
Table 3. Contribution rate of climate change and human activities to the runoff changes in the Dongyanghe River Basin, Nanyanghe River Basin, Yanghe River Basin based on the random forest model (1986−2015)
水文站
Hydrometric station时段
Time interval实测径流
Observed
runoff
(×108 m3)模拟径流
Simulated
runoff
(×108 m3)径流变化量
Runoff
variation
(×108 m3)贡献值
Variable quantity (×108 m3)贡献率
Contribution rate (%)气候变化
Climate change人类活动
Human activities气候变化
Climate change人类活动
Human activities柴沟堡东
Chaigoubu East Station突变前
Before abrupt change0.93 突变后
After abrupt change0.42 0.87 −0.51 −0.06 −0.45 11.5 88.5 柴沟堡南
Chaigoubu South Station突变前
Before abrupt change0.47 突变后
After abrupt change0.19 0.46 −0.28 −0.01 −0.27 2.4 97.6 响水堡
Xiangshuibu Station突变前
Before abrupt change2.14 突变后
After abrupt change0.95 2.10 −1.20 −0.04 −1.15 3.4 96.6 -
[1] LIU C M, XIA J. Water problems and hydrological research in the Yellow River and the Huai and Hai River basins of China[J]. Hydrological Processes, 2004, 18(12): 2197−2210 doi: 10.1002/hyp.5524 [2] MONTANARI A, YOUNG G, SAVENIJE H H G, et al. “Panta Rhei—Everything Flows”: Change in hydrology and society — the IAHS scientific decade 2013−2022[J]. Hydrological Sciences Journal, 2013, 58(6): 1256−1275 doi: 10.1080/02626667.2013.809088 [3] 徐小元. 永定河综合治理与生态修复上升为国家战略谋划[J]. 中国水利, 2019(22): 45−46 doi: 10.3969/j.issn.1000-1123.2019.22.026XU X Y. Strategic planning for rising restoration and improvement of ecological function zone of the Yongding River to the national strategy[J]. China Water Resources, 2019(22): 45−46 doi: 10.3969/j.issn.1000-1123.2019.22.026 [4] 侯蕾. 北方水资源短缺流域生态-水文响应机制研究——以永定河为例[D]. 北京: 中国水利水电科学研究院, 2019HOU L. Study on mechanism of ecohydrological response at the water resources shortage watershed in Northern China — A case study of Yongding River[D]. Beijing: China Institute of Water Resources and Hydropower Research, 2019 [5] 慕星, 赵勇, 刘欢, 等. 气候变化和人类活动影响下径流演变研究进展[J]. 人民黄河, 2021, 43(5): 35−41 doi: 10.3969/j.issn.1000-1379.2021.05.007MU X, ZHAO Y, LIU H, et al. Research advances on impacts of climate change and human activities on streamflow variation[J]. Yellow River, 2021, 43(5): 35−41 doi: 10.3969/j.issn.1000-1379.2021.05.007 [6] 李凌程, 张利平, 夏军, 等. 气候波动和人类活动对南水北调中线工程典型流域径流影响的定量评估[J]. 气候变化研究进展, 2014, 10(2): 118−126LI L C, ZHANG L P, XIA J, et al. Quantitative assessment of impacts of climate variability and human activities on runoff change in the typical basin of the middle route of the south-to-north water transfer project[J]. Climate Change Research, 2014, 10(2): 118−126 [7] HU S S, LIU C M, ZHENG H X, et al. Assessing the impacts of climate variability and human activities on streamflow in the water source area of Baiyangdian Lake[J]. Journal of Geographical Sciences, 2012, 22(5): 895−905 doi: 10.1007/s11442-012-0971-9 [8] 贾仰文. 分布式流域水文模型原理与实践[M]. 北京: 中国水利水电出版社, 2005JIA Y W. Principle and Practice of Distributed Watershed Hydrological Model[M]. Beijing: China Water Power Press, 2005 [9] 夏军, 马协一, 邹磊, 等. 气候变化和人类活动对汉江上游径流变化影响的定量研究[J]. 南水北调与水利科技, 2017, 15(1): 1−6XIA J, MA X Y, ZOU L, et al. Quantitative analysis of the effects of climate change and human activities on runoff in the upper Hanjiang River basin[J]. South-to-North Water Transfers and Water Science & Technology, 2017, 15(1): 1−6 [10] LI F P, ZHANG G X, XU Y. Separating the impacts of climate variation and human activities on runoff in the Songhua River basin, Northeast China[J]. Water, 2014, 6(11): 3320−3338 doi: 10.3390/w6113320 [11] ZHAO G J, TIAN P, MU X M, et al. Quantifying the impact of climate variability and human activities on streamflow in the middle reaches of the Yellow River Basin, China[J]. Journal of Hydrology, 2014, 519: 387−398 doi: 10.1016/j.jhydrol.2014.07.014 [12] WU J W, MIAO C Y, ZHANG X M, et al. Detecting the quantitative hydrological response to changes in climate and human activities[J]. Science of the Total Environment, 2017, 586: 328−337 doi: 10.1016/j.scitotenv.2017.02.010 [13] 夏润亮, 刘启兴, 李涛, 等. 基于集成学习的黄河未控区径流预测研究[J]. 应用基础与工程科学学报, 2020, 28(3): 740−749XIA R L, LIU Q X, LI T, et al. Research on runoff prediction of uncontrolled areas of the Yellow River based on ensemble learning[J]. Journal of Basic Science and Engineering, 2020, 28(3): 740−749 [14] 徐中民, 蓝永超, 程国栋. 人工神经网络方法在径流预报中的应用[J]. 冰川冻土, 2000, 22(4): 372−375XU Z M, LAN Y C, CHENG G D. A study on runoff forecast by aritifical neural network model[J]. Journal of Glaciolgy and Geocryology, 2000, 22(4): 372−375 [15] 王景雷, 吴景社, 孙景生, 等. 支持向量机在地下水位预报中的应用研究[J]. 水利学报, 2003, 34(5): 122−128 doi: 10.3321/j.issn:0559-9350.2003.05.022WANG J L, WU J S, SUN J S, et al. Application of support vector machine method in prediction of groundwater level[J]. Journal of Hydraulic Engineering, 2003, 34(5): 122−128 doi: 10.3321/j.issn:0559-9350.2003.05.022 [16] TYRALIS H, PAPACHARALAMPOUS G, LANGOUSIS A. A brief review of random forests for water scientists and practitioners and their recent history in water resources[J]. Water, 2019, 11(5): 910 doi: 10.3390/w11050910 [17] 李伶杰, 王银堂, 胡庆芳, 等. 基于随机森林与支持向量机的水库长期径流预报[J]. 水利水运工程学报, 2020(4): 33−40 doi: 10.12170/20190626001LI L J, WANG Y T, HU Q F, et al. Long-term inflow forecast of reservoir based on random forest and support vector machine[J]. Hydro-Science and Engineering, 2020(4): 33−40 doi: 10.12170/20190626001 [18] SHORTRIDGE J E, GUIKEMA S D, ZAITCHIK B F. Machine learning methods for empirical streamflow simulation: a comparison of model accuracy, interpretability, and uncertainty in seasonal watersheds[J]. Hydrology and Earth System Sciences, 2016, 20(7): 2611−2628 doi: 10.5194/hess-20-2611-2016 [19] PAPACHARALAMPOUS G A, TYRALIS H. Evaluation of random forests and prophet for daily streamflow forecasting[J]. Advances in Geosciences, 2018, 45: 201−208 doi: 10.5194/adgeo-45-201-2018 [20] LI B, YANG G S, WAN R R, et al. Comparison of random forests and other statistical methods for the prediction of lake water level: a case study of the Poyang Lake in China[J]. Hydrology Research, 2016, 47(S1): 69−83 doi: 10.2166/nh.2016.264 [21] KUPERS S J, WIRTH C, ENGELBRECHT B M J, et al. Dry season soil water potential maps of a 50 hectare tropical forest plot on Barro Colorado Island, Panama[J]. Scientific Data, 2019, 6: 63 doi: 10.1038/s41597-019-0072-z [22] SHIRI J. Improving the performance of the mass transfer-based reference evapotranspiration estimation approaches through a coupled wavelet-random forest methodology[J]. Journal of Hydrology, 2018, 561: 737−750 doi: 10.1016/j.jhydrol.2018.04.042 [23] JIANG B, WONG C P, LU F, et al. Drivers of drying on the Yongding River in Beijing[J]. Journal of Hydrology, 2014, 519: 69−79 doi: 10.1016/j.jhydrol.2014.06.033 [24] 张利平, 于松延, 段尧彬, 等. 气候变化和人类活动对永定河流域径流变化影响定量研究[J]. 气候变化研究进展, 2013, 9(6): 391−397ZHANG L P, YU S Y, DUAN Y B, et al. Quantitative assessment of the effects of climate change and human activities on runoff in the Yongding River Basin[J]. Climat Change Research, 2013, 9(6): 391−397 [25] 田菲, 韩淑敏, 胡玉昆. 洋河流域径流演变规律及驱动因子分析[J]. 华北农学报, 2008, 23(S2): 353−357 doi: 10.7668/hbnxb.2008.S2.081TIAN F, HAN S M, HU Y K. The evolution trend and motivating factors of runoff in Yang River Basin[J]. Acta Agriculturae Boreali-Sinica, 2008, 23(S2): 353−357 doi: 10.7668/hbnxb.2008.S2.081 [26] XIA J, ZENG S D, DU H, et al. Quantifying the effects of climate change and human activities on runoff in the water source area of Beijing, China[J]. Hydrological Sciences Journal, 2014, 59(10): 1794−1807 doi: 10.1080/02626667.2014.952237 [27] 王绍瑛. 永定河的治理成就与存在问题[J]. 北京水利, 1997(3): 25−27WANG S Y. Achievements and existing problems in the teatment of Yongding River[J]. Beijing Water Resources, 1997(3): 25−27 [28] 侯蕾, 彭文启, 董飞, 等. 永定河上游流域水文气象要素的历史演变特征[J]. 中国农村水利水电, 2020(12): 1−8, 14 doi: 10.3969/j.issn.1007-2284.2020.12.001HOU L, PENG W Q, DONG F, et al. Historical evolution of the hydro-meteorological elements in the upstream of the Yongding River Basin[J]. China Rural Water and Hydropower, 2020(12): 1−8, 14 doi: 10.3969/j.issn.1007-2284.2020.12.001 [29] 魏凤英. 现代气候统计诊断与预测技术[M]. 2版. 北京: 气象出版社, 2007: 1–296WEI F Y. Modern Meteorological Statistical Diagnosis and Prediction Technology[M]. Beijing: China Meteorological Press, 2007: 1–296 [30] MCDONALD S, MOHAMMED I N, BOLTEN J D, et al. Web-based decision support system tools: The Soil and Water Assessment Tool Online visualization and analyses (SWATOnline) and NASA earth observation data downloading and reformatting tool (NASAaccess)[J]. Environmental Modelling & Software, 2019, 120: 104499 [31] 黄奎. SWAT模型研究进展[J]. 珠江水运, 2019(19): 34−35HUANG K. Research progress of SWAT model[J]. Pearl River Water Transport, 2019(19): 34−35 [32] 马骊. 随机森林算法的优化改进研究[D]. 广州: 暨南大学, 2016MA L. Research on optimization and improvement of random forests algorithm[D]. Guangzhou: Jinan University, 2016 [33] 秦英. 基于随机森林的WebShell检测方法[J]. 计算机系统应用, 2019, 28(2): 240−245QIN Y. Webshell detection method based on random forest[J]. Computer Systems & Applications, 2019, 28(2): 240−245 [34] 张建云, 王国庆. 河川径流变化及归因定量识别[M]. 北京: 科学出版社, 2014: 18–25ZHANG J Y, WANG G Q. Quantitative Identification of River Runoff Change and Attribution[M]. Beijing: Science Press, 2014: 18–25 [35] 李秀. 永定河流域径流演变特征及驱动因素分析[D]. 沈阳: 沈阳大学, 2021LI X. Analysis of runoff evolution characteristics and driving factors in Yongding River Basin[D]. Shenyang: Shenyang University, 2021 [36] 王艺璇, 沈彦军, 高雅, 等. 永定河上游环境变化和水资源演变研究进展[J]. 南水北调与水利科技(中英文), 2021, 19(4): 656−668WANG Y X, SHEN Y J, GAO Y, et al. Research progress on the changes of environment and water resources in the upper Yongding River Basin[J]. South-to-North Water Transfers and Water Science & Technology, 2021, 19(4): 656−668 [37] 张雷, 王琳琳, 张旭东, 等. 随机森林算法基本思想及其在生态学中的应用−以云南松分布模拟为例[J]. 生态学报, 2014, 34(3): 650−659ZHANG L, WANG L L, ZHANG X D, et al. The basic principle of random forest and its applications in ecology: a case study of Pinus yunnanensis[J]. Acta Ecologica Sinica, 2014, 34(3): 650−659 [38] 宋晓猛, 张建云, 占车生, 等. 气候变化和人类活动对水文循环影响研究进展[J]. 水利学报, 2013, 44(7): 779−790SONG X M, ZHANG J Y, ZHAN C S, et al. Review for impacts of climate change and human activities on water cycle[J]. Journal of Hydraulic Engineering, 2013, 44(7): 779−790 [39] 苏辉东, 贾仰文, 倪广恒, 等. 机器学习在径流预测中的应用研究[J]. 中国农村水利水电, 2018(6): 40−43, 48 doi: 10.3969/j.issn.1007-2284.2018.06.009SU H D, JIA Y W, NI G H, et al. The application of machine learning in runoff prediction[J]. China Rural Water and Hydropower, 2018(6): 40−43, 48 doi: 10.3969/j.issn.1007-2284.2018.06.009 -