Influence of Quantity and Arrangement of Rain Gauges on Accuracy of Runoff Simulation in Medium-sized Watershed in Karst Areas: Evidence from Dumu River Basin
Runoff simulation is greatly impacted by the quantity and arrangement of rain gauges, especially in difficult terrain like the southwest mountainous area where there are frequent flash floods and unevenly distributed rain gauges. The Dumu River Basin, a common watershed in the upper levels of the Wujiang River that is vulnerable to flash floods, was chosen as the subject region for this investigation. The semi-distributed hydrological model TOPMODEL, which is easy to integrate and carry out a large number of stochastic simulations, was selected as the carrier for the rainfall and runoff observation data in the basin. This study quantitatively demonstrates the impact of the number and distribution characteristics of rain gauges on the simulation accuracy of typical karst basins using a variety of error analysis indicators and rain gage network spatial distribution description indicators. According to the findings, the range shows a negative exponential square decay, whereas the mean, standard deviation, coefficient of variation, minimum value, and maximum value all change exponentially. The maximum values of the NSE and correlation coefficient stay comparatively constant, suggesting that by optimizing the network distribution of rain gauges, great simulation accuracy can be attained even with a small number of rain gauges. All hydrological model indicators show a quadratic function relationship with the nearest neighbor index (NNI). Information entropy shows the same trend as correlation coefficients and NSE with regard to the number of rain gauges, and it has a greater correlation with error indicators than NNI. The maximum model accuracy is achieved when NNI falls between 1 and 2. A certain amount of uncertainty is present in both NNI and information entropy: the same value may accompany multiple simulation accuracies. This suggests that while establishing a rain gage network with a single indicator can effectively analyze areal rainfall, it may not always accurately simulate runoff. The results of the study can serve as technical references for water resource assessment, flood analysis and forecasting, and network optimization for regional rain gauges.
ZENG Shi-jie, ZHOU Xiang-yang, CHEN Bai-lian, LEI Wen-juan, ZHONG Lei-lei, XIAO Gui-hong.
Influence of Quantity and Arrangement of Rain Gauges on Accuracy of Runoff Simulation in Medium-sized Watershed in Karst Areas: Evidence from Dumu River Basin[J].
Science Technology and Engineering, 2026, 26(13): 5348-5362 DOI:10.12404/j.issn.1671-1815.2504358
研究表明,流域布设雨量站的数量和空间分布均对模拟精度产生重要影响。雨量站数量的增加可明显提升模拟精度[4-5],二者的关系表现为非线性:当站网密度超过某一临界值时,模型精度的增加率提升较小[6]。需要注意的是,降雨的时空差异变化大[7]。雨量站空间分布方面,其分布的均匀程度也显著影响模型的影响结果。通常表现为:当雨量站数目较少时,雨量站空间分布不均匀对面平均雨量和径流模拟结果影响显著[8-9]。分布的均匀程度可利用信息熵进行衡量,表现出比其他方法的显著性更强,应用前景更好[10]。最大熵准则(principle of maximum entropy,POME)也被应用于评估雨量站网的分布[11],边际熵与总降水量呈现显著相关性,降雨丰沛区域的概率分布也更均匀[12]。但上述研究揭示雨量站数量与模型精度的关系时指标不够系统,非线性的演变关系也极少量化。
选取不同站网密度的典型代表:低密度(2个站点)、中低密度(4个站点)、中等密度(8个站点)和高密度(14个站点),并识别各等级站网分布最不均匀、中等均匀和最均匀3种情形,对应NNI最小值、中位数和最大值。上述各情形的站点的空间分布如图7所示。可以看出,无论在哪种雨量站密度下,NNI 指数最大时,其得到的雨量站在整个研究区空间分布均匀;而 NNI 最小时,其得到的雨量站在整个研究区空间分布聚集。
XuJing, RenLiliang, LiuXiaofan, et al. Impact of rainfall uncertainty propagation based on fuzzy set theory[J]. Advances in Water Science, 2009, 20(3): 422-427.
MiJiawei, TianJiyang, XueHai, et al. Evaluation of the application of radar rainfall measurement and nowcasting from the hydrological per spective of small and medium basins[J]. Science Technology and Engineering, 2022, 22(23): 10247-10254.
[5]
SunS, KhuS T, DjordjevicS. Sampling rainfall events: a novel approach to generate large correlated samples[J]. Hydrology Research, 2013, 44: 351-361.
GuoWeiguo, ChenXi, ZhangRunrun. Study on the applicability of spatial interpolation methods based on the heterogeneity of rainfall distribution[J]. Water Power, 2016, 42(6): 14-17.
[8]
AnctilF, LauzonN, AndréassianV, et al. Improvement of rainfall runoff forecasts through mean areal rainfall optimization[J]. Journal of Hydrology, 2006, 328(3): 717-725.
[9]
BardossyA, DasT. Influence of rainfall observation network on model calibration and application[J]. Hydrology & Earth System Sciences, 2006, 12(1): 77-89.
ChenAijun, TaoZichun. Comprehensive assessment of IMERG accuracy under diverse topographic conditions[J]. Science Technology and Engineering, 2025, 25(16): 6608-6618.
[12]
XuH L, XuC Y, ChenH, et al. Assessing the influence of rain gauge density and distribution on hydrological model performance in a humid region of China[J]. Journal of Hydrology, 2013, 505: 1-12.
[13]
XuH L, XuC Y, SaelthunN R, et al. Entropy theory based multi-criteria resampling of rain gauge networks for hydrological modelling: a case study of humid area in southern China[J]. Journal of Hydrology, 2015, 525: 138-151.
[14]
MishraA K, CoulibalyP. Developments in hydrometric network design: a review[J]. Reviews of Geophysics, 2009, 47(2): RG2001.
[15]
KrstanovicP F, SinghV P. Evaluation of rainfall networks using entropy: I. theoretical development[J]. Water Resources Management, 1992, 6(4): 279-293.
[16]
SilvaV P R, BeloF A F, SinghV P, et al. Entropy theory for analysing water resources in northeastern region of Brazil[J]. Hydrological Sciences Journal, 2017, 62(7): 1029-1038.
[17]
ZhengW, JiangJ, TaoK. A method based on musical-staff-inspired signal processing model for measuring rock moisture content[J]. Measurement, 2018, 125: 577-585.
LiangGuixing, QinXiaoqun, CuiYali, et al. Improvement and application of a distributed hydrological model in karst regions[J]. Hydrogeology & Engineering Geology, 2020, 47(2): 60-67.
[20]
DoummarJ, SauterM, GeyerT. Simulation of flow processes in a large scale karst system with an integrated catchment model (Mike She): identification of relevant parameters influencing spring discharge[J]. Journal of Hydrology, 2012, 426/427: 112-123.
[21]
BirkS, GeyerT, LiedlR, et al. Process-based interpretation of tracer tests in carbonate aquifers[J]. Ground Water, 2005, 43(3): 381-388.
WangYujie, LiZhijia, YaoCheng, et al. Flood simulation in the Qinhuai River Basin based on a coupled distributed hydrological and hydrodynamic model[J]. Journal of Hohai University(Natural Sciences), 2025, 53(2): 11-18, 114.
[24]
XuC, XuX, LiuM, et al. An improved optimization scheme for representing hillslopes and depressions in karst hydrology[J]. Water Resources Research, 2020, 56(5): 2019WR026038.
ChenHua, HuoRan, ZengQiang, et al. Comparative study on the influence of rain-gauge network on the uncertainty of hydrological modeling[J]. Advances in Water Science, 2019, 30(1): 34-44.
[27]
YinH, ZhaoL, ZhuM, et al. Runoff prediction in gauged and ungauged basins using Transformer-XAJ model[J]. Journal of Hydrology, 2025, 662: 133954.
[28]
陈石磊. 西南典型喀斯特流域分布式降雨径流模拟研究[D]. 武汉: 武汉大学, 2020.
[29]
ChenShilei. Study on distributed rainfall-runoff simulation in a typical karst catchment, southwest China[D]. Wuhan: Wuhan University, 2020.
ChenJingkai, ZhangZhicai, ChenXi, et al. Simulation analysis of rainfall-runoff processes and their responses to wet-dry condition variations in a karst watershed[J]. Journal of Hohai University (Natural Sciences), 2025, 53(1): 41-49.
[32]
FordD, WilliamsP. Karst hydrogeology and geomorphology[M]. Hoboken: Wiley, 2007.
MaCongwen, ZhangZhicai, ChenXi, et al. Simulation of karst spring discharge in southwest China using machine learning approaches[J]. Carsologica Sinica, 2024, 43(1): 48-56.
YangShuping, HeZhonghua, WangMaoqiang, et al. Study on rainfall-runoff mechanisms of medium structure in typical karst basin-taking the central Guizhou region as an example[J]. Research of Soil and Water Conservation, 2024, 31(3): 49-60.
ZhangJun, ChenHongsong, NieYunpeng, et al. Research progress on structure and hydrological processes in the karst critical zone of southwest China[J]. Chinese Journal of Applied Ecology, 2024, 35(4): 985-996.
[39]
ZhangZ, ChenX, ChengQ, et al. Storage dynamics, hydrological connectivity and flux ages in a karst catchment: conceptual modelling using stable isotopes[J]. Hydrology and Earth System Sciences, 2019, 23(1): 51-71.
[40]
BevenK J, KirkbyM J. A physically based, variable contributing area model of basin hydrology[J]. Hydrological Sciences Bulletin, 1979, 24(1): 43-69.
[41]
QiuA N, ZhangY J, WangG X, et al. A modified TOPMODEL introducing the bedrock surface topographic index in Huangbengliu watershed, China[J]. Journal of Mountain Science, 2022, 19(12): 3517-3532.
JinQianfang, ZhouJianzhong, YanDong, et al. Hydrological simulation of TOPMODEL model in ungauged basin of the upper reaches of the Yangtze River[J]. Water Resources and Power, 2018, 36(9): 44-47.
[44]
HaoG, LiJ, SongL, et al. Comparison between the TOPMODEL and the Xin’anjiang model and their application to rainfall runoff simulation in semi-humid regions[J]. Environmental Earth Sciences, 2018, 77(7): 1-13.
[45]
LiW, LinK, ZhaoT, et al. Risk assessment and sensitivity analysis of flash floods in ungauged basins using coupled hydrologic and hydrodynamic models[J]. Journal of Hydrology, 2019, 572: 108-120.
[46]
BevenK J, KirkbyM J, FreerJ E, et al. A history of TOPMODEL[J]. Hydrology and Earth System Sciences, 2021, 25(2): 527-549.
WangMengmeng, WeiZhanmin, ZhangRuiqiang, et al. Quantitative evaluation of basic flow segmentation results in Beiluo River based on TOPMODEL hydrological model[J]. Yellow River, 2020, 42(4): 19-24.
[49]
JeziorskaJ, NiedzielskiT. Applicability of TOPMODEL in the mountainous catchments in the upper Nysa Kłodzka River Basin (SW Poland)[J]. Acta Geophysica, 2018, 66(2): 203-222.
[50]
GumindogaW, RwasokaD T, MurwiraA. Simulation of streamflow using TOPMODEL in the Upper Save River catchment of Zimbabwe[J]. Physics and Chemistry of the Earth, Parts A/B/C, 2011, 36(14/15): 806-813.
LiangRuhao, LinKairong, LinYouqin, et al. Application of TOPMODEL-based distributed flood forecasting model for small-sized and medium-sized rivers[J]. Water Resources and Hydropower Engineering, 2019, 50(9): 62-68.
[53]
熊立华, 郭生练. 分布式流域水文模型[M]. 北京: 中国水利水电出版社, 2004.
[54]
XiongLihua, GuoShenglian. Distributed watershed hydrological model[M]. Beijing: China Water & Power Press, 2004.
[55]
QuinnP, BevenK J, LambR. The ln(a/tanβ) index: how to calculate it and how to use it in the TOPMODEL framework[J]. Hydrological Processes, 1995, 9(2): 161.
[56]
QiuA N, ZhangY J, WangG X, et al. A modified TOPMODEL introducing the bedrock surface topographic index in Huangbengliu watershed, China[J]. Journal of Mountain Science, 2022, 19(12): 3517-3532.
FuLiqiang, KongShi, ZongCheng, et al. The difference of spatial distribution of wetland nature reserves and wetland parks in China[J]. Wetland Science, 2015, 13(3): 356-363.
[59]
LeeJ H, JunH D. Influence of the spatial distribution of a raingauge network on the estimation of areal average rainfall: focusing on Thiessen’s weighting method[J]. Journal of Korean Society of Hazard Mitigation, 2015, 15(3): 297-306.
XiaZiyang, XiaYunfan, WangNing, et al. Spatial pattern and influencing factors of tourism elements in urban agglomeration on northern slope of Tianshan Mountains based on POI data[J]. Arid Land Geography, 2025, 48(7): 1243-1254.
[62]
ShannonC E. A mathematical theory of communication[J]. The Bell System Technical Journal, 1948, 27(3): 379-423.
[63]
LiC, SinghV P, MishraA K. Entropy theory-based criterion for hydrometric network evaluation and design: maximum information minimum redundancy[J]. Water Resources Research, 2012, 48: W05521.
ZhangJinnan, WuJian, WeiGuozhen, et al. Method of flood forecasting based on rainfall input improvement by optimizing rainfall stations’ weights[J]. Journal of Dalian University of Technology, 2020, 60(5): 537-546.
LiuZhiwu, LiZiheng, SangXuefeng, et al. Spatio-temporal rainfall distribution method based on synergistic matrix coupling coupling Thiessen polygons: a case study in the Yangtze River Basin[J]. Journal of China Hydrology, 2023, 43(5): 51-56.
[70]
MongerF, SpracklenD V, KirkbyM J, et al. Investigating the impact of woodland placement and percentage cover on flood peaks in an upland catchment using spatially distributed TOPMODEL[J]. Journal of Flood Risk Management, 2024, 17(2): 12977.
YangWushuang, HuoRan, ZengQiang, et al. An analysis of hydrological simulation effect under different rain gauge density[J]. China Rural Water and Hydropower, 2020(5): 42-46, 52.
[73]
RossS M.Introduction to probability models[M]. 5th Ed. Santiago: Academic Press, 2015.
[74]
SaqibM, MemonZ A. Some remarks on extreme sample order statistics relating to power Lomax geometric distribution[J]. Scientific African, 2025, 30: e03008.
WangQingjing, YangXinyue, ChenHua, et al. Influence of rain gauges network configuration on the accuracy of rainfall spatial interpolation and hydrological modeling[J]. Journal of Yangtze River Scientific Research Institute, 2019, 36(4): 19-26.
[77]
LiD, LiP, HuJ, et al. A DEM differencing method for detecting geomorphic changes on topographically complex areas based on RAV remote sensing techniques[J]. IEEE Transactions on Geoscience and Remote Sensing, 2025, 63: 1-15.
LiuBo, NiuQuanfu, WangGang, et al. Analysis of forest cover changes and driving forces in the Loess Plateau (Gansu region) based on multisensor remote sensing images[J]. Science Technology and Engineering, 2025, 25(1): 54-66.
[80]
DobreB, IstvánP K, BugyaT. Comparison of digital elevation models through the analysis of geomorphic surface remnants in the Desatoya Mountains, Nevada[J]. Transactions in GIS, 2021, 25(5): 2262-2282.
ChiHongqing, PengXiaodi, HuXianzhen, et al. Quantitative analysis of the optimal extraction area of relief amplitude and geomorphological classification research: a case study of Wuan City[J]. Science Technology and Engineering, 2025, 25(10): 4017-4026.
DingWenfeng, HanHaoyu, WangXiekang, et al. Rainfall networks optimization in small catchment of loess plateau based on information entropy[J]. Advanced Engineering Sciences, 2022, 54(2): 113-121.