Prediction of lithospheric heat flow of the South China Sea Oceanic crust based on machine learning methods

doi:10.13745/j.esf.sf.2025.4.60

Abstract

Abstract:

Heat flow is a critical parameter in geodynamic studies and resource exploration, but its measurements are often susceptible to interference from factors such as climate and hydrothermal activity, resulting in scarce observational data. This study addresses the insufficient consideration of oceanic crust specificity in existing machine learning-based heat flow prediction models. By integrating measured heat flow and multi-source geological and geophysical data from the South China Sea, we employ Linear Model, Support Vector Machine, and XGBoost algorithms to compare prediction models with and without oceanic crust features (distance to mid-ocean ridge and oceanic crust age), revealing the influence mechanisms of oceanic crust characteristics on the regional heat flow distribution. The results show that oceanic crust features exhibit no significant correlation with measured heat flow values, yet they result in a more pronounced zonal distribution of predicted heat flow near the mid-ocean ridge. In models incorporating these features, the feature importance of Bouguer gravity anomaly becomes significantly higher than that of other features. Furthermore, K-means clustering analysis based on heat flow values, Bouguer gravity anomaly, and distance to mid-ocean ridge identifies two distinct oceanic crust types: tectonically dominated (Type 1) and magmatically dominated (Type 2) regions. This supports the evolutionary trend of the South China Sea’s oceanic crust spreading mechanism shifting from magmatically to tectonically dominated after the mid-ocean ridge jump at 23.6 Ma. This study provides a new data-driven framework for understanding the deep dynamic processes of the South China Sea.

Key words: geothermal heat flow, South China Sea, machine learning, oceanic crust, mid-ocean ridge

CLC Number:

P314.3

ZHANG Yufei, ZHANG Yang, JI Junjie, CHENG Qiuming. Prediction of lithospheric heat flow of the South China Sea Oceanic crust based on machine learning methods[J]. Earth Science Frontiers, 2025, 32(4): 235-249.

Figures/Tables 16

Fig.1 Topography of the South China Sea and division of tectonic blocks. Terrain data according to reference [24]; Boundary data of tectonic units, microplate boundaries of the South China Sea, and tectonic blocks of the South China Sea are based on reference [25]; data for mid-ocean ridges 1, 2, and 3 are from reference [26].

Fig.2 The study area and measured heat flow points

Fig.3 Histogram of the frequency distribution of heat flow measurements

Table 1 Statistics of measured heat flow in various ocean basins in the South China Sea

各次级洋盆	数量/个	最大值/(mW·m^-2)	最小值/(mW·m^-2)	平均值/(mW·m^-2)	标准值/(mW·m^-2)	中位值/(mW·m^-2)
西北次海盆	10	115.2	8.2	72.8	37.6	86.1
东部次海盆	30	152.0	17.2	88.4	32.6	91.5
西南次海盆	31	152.0	11.1	99.5	27.8	99.5

Table 2 Characterization and its source datasets for the prediction of heat flow

编号	数据名称	数据来源
1	沉积物厚度	Straume 等^[47]
2	海底测深	Tozer 等^[24]
3	布格重力异常	Balmino等^[48]
4	均衡重力异常	Balmino等^[48]
5	自由空气重力异常	Balmino等^[48]
6	磁异常数据	Choi 等^[49]
7	岩石圈厚度	Hoggard等^[50]
8	结晶地壳(无沉积物)厚度	Mooney等^[51]
9	垂向重力梯度	Sandwell等^[52]
10	居里点深度	Li等^[39]
11	地形崎岖度
12	距海山距离	Yesson 等^[53]
13	距山丘距离	Yesson 等^[53]
14	洋壳年龄	Seton等^[54]
15	距洋中脊距离

Fig.4 Correlation between heat flow measurements and individual characteristics

Fig.5 Correlation between the measured values of heat flow and the characteristics of each basin

Fig.6 Evaluation of overall forecasting results

Fig.7 Scatterplot of actual versus and predicted values for the XGBoost

Fig.8 Predicted heat flow map of the South China Sea oceanic crust

Fig.9 Validation model for randomized hyperparameter training

Fig.10 Ranking of feature importance

Fig.11 3D spatial results of K-means clustering

Fig.12 Exploring the relationships between elements in Type 1 and Type 2 regional clusters

Fig.13 Spatial distribution of Type 1 and Type 2 regions

Fig.14 Correlation diagram between different categories of heat flux values and various features

References 62

[1]	刘峰, 王贵玲, 姜光政, 等. 我国陆区大地热流测量新进展与新认识[J]. 地学前缘, 2024, 31(6): 19-30. DOI
[2]	KIM Y G, LEE S M, JIN Y K, et al. The stability of gas hydrate field in the northeastern continental slope of Sakhalin Island, Sea of Okhotsk, as inferred from analysis of heat flow data and its implications for slope failures[J]. Marine and Petroleum Geology, 2013, 45: 198-207.
[3]	GAO X, WANG K L. Strength of stick-slip and creeping subduction megathrusts from heat flow observations[J]. Science, 2014, 345(6200): 1038-1041. DOI PMID
[4]	HENDERSON P. Annual review of earth and planetary sciences[J]. Geochimica et Cosmochimica Acta, 1988, 52(4): 950.
[5]	刘金萍, 王改云, 简晓玲, 等. 北黄海东部次盆地构造热机制与成烃效应[J]. 地学前缘, 2024, 31(4): 206-218. DOI
[6]	LUCAZEAU F. Analysis and mapping of an updated terrestrial heat flow data set[J]. Geochemistry, Geophysics, Geosystems, 2019, 20(8): 4001-4024. DOI
[7]	GRAW J H, WOOD W T, PHRAMPUS B J. Predicting marine in situ heat flow using a geospatial machine learning conformal prediction[J]. Geochemistry, Geophysics, Geosystems, 2023, 24(6): e2023GC010913.
[8]	LÖSING M, EBBING J. Predicting geothermal heat flow in Antarctica with a machine learning approach[J]. Journal of Geophysical Research: Solid Earth, 2021, 126(6): e2020JB021499.
[9]	HE J F, LI K W, WANG X W, et al. A machine learning methodology for predicting geothermal heat flow in the Bohai Bay basin, China[J]. Natural Resources Research, 2022, 31(1): 237-260.
[10]	周永章, 张良均, 张奥多, 等. 地球科学大数据挖掘与机器学习[M]. 广州: 中山大学出版社, 2018.
[11]	BELL G, HEY T, SZALAY A. Beyond the data deluge[J]. Science, 2009, 323(5919): 1297-1298.
[12]	蒋德鑫, 姜鹍鹏, 张贺, 等. 西北太平洋边缘海热流特征研究[J]. 现代地质, 2020, 34(1): 117-129.
[13]	王淑杰, 翟世奎, 于增慧, 等. 关于现代海底热液活动系统模式的思考[J]. 地球科学, 2018, 43(3): 835-850.
[14]	CORTES C, VAPNIK V. Support-vector networks[J]. Machine Learning, 1995, (20): 273-297.
[15]	CHEN T Q, GUESTRIN C. XGBoost: a scalable tree boosting system[C]// Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco: ACM, 2016: 785-794.
[16]	MACQUEEN J B. Some methods for classification and analysis of multivariate observations[J]. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1: Statistics, 1966(5): 281-298.
[17]	栾锡武, 张亮. 南海构造演化模式: 综合作用下的被动扩张[J]. 海洋地质与第四纪地质, 2009, 29(6): 59-74.
[18]	LI C F, WANG P X, DIETER F, et al. Opening of the South China Sea and its implications for southeast Asian tectonics since the late Mesozoic[J]. Advances in Earth Science, 2009, 24(12): 1339-1351.
[19]	施小斌, 丘学林, 夏戡原, 等. 南海热流特征及其构造意义[J]. 热带海洋学报, 2003, 22(2): 63-73.
[20]	BURTON-JOHNSON A, CULLEN A B. Continental rifting in the South China Sea through extension and high heat flow: an extended history[J]. Gondwana Research, 2023, 120: 235-263.
[21]	王鹏程, 李三忠, 郭玲莉, 等. 南海打开模式: 右行走滑拉分与古南海俯冲拖曳[J]. 地学前缘, 2017, 24(4): 294-319. DOI
[22]	ZHANG F, LIN J, ZHANG X B, et al. Asymmetry in oceanic crustal structure of the South China Sea basin and its implications on mantle geodynamics[J]. International Geology Review, 2020, 62(7/8): 840-858.
[23]	谢玉洪, 郑华安, 梁玉凯, 等. 南海盆地地热地质条件和资源形成因素[J]. 成都理工大学学报(自然科学版), 2023, 50(6): 647-660.
[24]	TOZER B, SANDWELL D T, SMITH W H F, et al. Global bathymetry and topography at 15 arc sec: SRTM15+[J]. Earth and Space Science, 2019, 6(10): 1847-1864. DOI
[25]	邱燕, 杜文波, 鞠东, 等. 南海主要断裂(带)与构造分区[J]. 海洋地质与第四纪地质, 2025, 45(1): 136-153.
[26]	梁光河. 南海中央海盆高精度地震勘探揭示的大陆漂移过程[J]. 地学前缘, 2022, 29(4): 293-306. DOI
[27]	FUCHS S, NORDEN B, NEUMANN F, et al. The Global Heat Flow Database: new quality scheme and data release 2023[C]// XXVIII General Assembly of the International Union of Geodesy and Geophysics. Berlin:IUGG, 2023.
[28]	徐行, 姚永坚, 彭登, 等. 南海西南次海盆的地热流特征与分析[J]. 地球物理学报, 2018, 61(7): 2915-2925. DOI
[29]	徐行, 董淼, 陈爱华, 等. 南海海盆IODP349钻井岩心的生热元素测试与应用研究[J]. 大地构造与成矿学, 2017, 41(6): 1128-1134.
[30]	NISSEN S S, HAYES D E, YAO B C, et al. Gravity, heat flow, and seismic constraints on the processes of crustal extension: northern margin of the South China Sea[J]. Journal of Geophysical Research: Solid Earth, 1995, 100(B11): 22447-22483.
[31]	TAYLOR B, HAYES D E. Origin and history of the South China Sea basin[J]. Geophysical Monograph Series, 1983, 27: 23-56.
[32]	HU L T, ZHANG P Z, HAO T Y, et al. Analysis of topography, heat flow against crustal age in the Southwest Subbasin, South China Sea[J]. International Geology Review, 2020, 62(7/8): 1081-1093.
[33]	HASTEROK D, CHAPMAN D S, DAVIS E E. Oceanic heat flow: implications for global heat loss[J]. Earth and Planetary Science Letters, 2011, 311(3/4): 386-395.
[34]	施小斌, 于传海, 陈梅, 等. 南海北部陆缘热流变化特征及其影响因素分析[J]. 地学前缘, 2017, 24(3): 56-64. DOI
[35]	SCLATER J G, JAUPART C, GALSON D. The heat flow through oceanic and continental crust and the heat loss of the Earth[J]. Reviews of Geophysics, 1980, 18(1): 269-311.
[36]	POLLACK H N, HURTER S J, JOHNSON J R. Heat flow from the earth’s interior: analysis of the global data set[J]. Reviews of Geophysics, 1993, 31(3): 267-280.
[37]	KUTAS R I. Heat flow, radiogenic heat and crustal thickness in southwest U.S.S.R[J]. Tectonophysics, 1984, 103(1/2/3/4): 167-174.
[38]	FISHER A T, BECKER K. Channelized fluid flow in oceanic crust reconciles heat-flow and permeability data[J]. Nature, 2000, 403(6765): 71-74.
[39]	LI C F, LU Y, WANG J. A global reference model of Curie-point depths based on EMAG2[J]. Scientific Reports, 2017, 7: 45129.
[40]	CHENG Q M. Singularity of lithosphere mass density over the mid-ocean ridges and implication on floor depth and heat flow[J]. Geoscience Frontiers, 2023, 14(5): 101591.
[41]	RÓŻYCKA M, MIGOŃ P, MICHNIEWICZ A. Topographic Wetness Index and Terrain Ruggedness Index in geomorphic characterisation of landslide terrains, on examples from the Sudetes, SW Poland[J]. Zeitschrift für Geomorphologie Supplement, 2017, 61(2): 61-80.
[42]	TONTINI F C, TIVEY M A, DE RONDE C E J, et al. Heat flow and near-seafloor magnetic anomalies highlight hydrothermal circulation at brothers volcano caldera, southern Kermadec arc, New Zealand[J]. Geophysical Research Letters, 2019, 46(14): 8252-8260.
[43]	吴招才, 高金耀, 沈中延, 等. 中国东部及近海磁异常特征及大地构造意义[J]. 地学前缘, 2018, 25(1): 210-217. DOI
[44]	COCHRAN J R, TALWANI M. Free-air gravity anomalies in the world’s oceans and their relationship to residual elevation[J]. Geophysical Journal International, 1977, 50(3): 495-552.
[45]	殷秀华, 刘铁胜, 刘占坡. 均衡重力异常和地壳表、浅层地质结构[J]. 地震地质, 1993, 15(2): 149-156.
[46]	HARPER H, TOZER B, SANDWELL D T, et al. Marine vertical gravity gradients reveal the global distribution and tectonic significance of “seesaw” ridge propagation[J]. Journal of Geophysical Research: Solid Earth, 2021, 126(2): e2020JB020017.
[47]	STRAUME E O, GAINA C, MEDVEDEV S, et al. GlobSed: updated total sediment thickness in the world’s oceans[J]. Geochemistry, Geophysics, Geosystems, 2019, 20(4): 1756-1772.
[48]	BALMINO G, VALES N, BONVALOT S, et al. Spherical harmonic modelling to ultra-high degree of Bouguer and isostatic anomalies[J]. Journal of Geodesy, 2012, 86(7): 499-520.
[49]	DYMENT J, CHOI Y, LESUR V, et al. The world digital magnetic anomaly map: version 2.1[C]// EGU General Assembly 2020, Online, 2020: EGU 2020-22098.
[50]	HOGGARD M J, CZARNOTA K, RICHARDS F D, et al. Global distribution of sediment-hosted metals controlled by craton edge stability[J]. Nature Geoscience, 2020, 13: 504-510.
[51]	MOONEY W D, BARRERA-LOPEZ C, SUÁREZ M G, et al. Earth crustal model 1 (ECM1): a 1° x 1° global seismic and density model[J]. Earth-Science Reviews, 2023, 243: 104493.
[52]	SANDWELL D T, MÜLLER R D, SMITH W H F, et al. New global marine gravity model from CryoSat-2 and Jason-1 reveals buried tectonic structure[J]. Science, 2014, 346(6205): 65-67.
[53]	YESSON C, CLARK M R, TAYLOR M L, et al. The global distribution of seamounts based on 30 arc seconds bathymetry data[J]. Deep Sea Research Part I: Oceanographic Research Papers, 2011, 58(4): 442-453.
[54]	SETON M, MÜLLER R D, ZAHIROVIC S, et al. A global data set of present-day oceanic crustal age and seafloor spreading parameters[J]. Geochemistry, Geophysics, Geosystems, 2020, 21(10): e2020GC009214.
[55]	PETRELLI M, PERUGINI D. Solving petrological problems through machine learning: the study case of tectonic discrimination using geochemical and isotopic data[J]. Contributions to Mineralogy and Petrology, 2016, 171(10): 81.
[56]	KOHAVI R. A study of cross-validation and bootstrap for accuracy estimation and model selection[J]. International Joint Conference on Artificial Intelligence, 1995, 2: 1137-1143.
[57]	DERBELA M, NOUIRI I. Intelligent approach to predict future groundwater level based on artificial neural networks (ANN)[J]. Euro-Mediterranean Journal for Environmental Integration, 2020, 5(3): 51.
[58]	黄磊, 陈泓君, 高红芳, 等. 南海中央海盆热流特征及成因[J]. 海洋地质前沿, 2013, 29(11): 39-43.
[59]	CURRIE C A, HYNDMAN R D. The thermal structure of subduction zone back arcs[J]. Journal of Geophysical Research: Solid Earth, 2006, 111(B8): B08404.
[60]	CANNAT M, SAUTER D, MENDEL V, et al. Modes of seafloor generation at a melt-poor ultraslow-spreading ridge[J]. Geology, 2006, 34(7): 605.
[61]	KAWADA Y, YAMANO M, SEAMA N. Hydrothermal heat mining in an incoming oceanic plate due to aquifer thickening: explaining the high heat flow anomaly observed around the Japan Trench[J]. Geochemistry, Geophysics, Geosystems, 2014, 15(4): 1580-1599.
[62]	MUTTER J C, KARSON J A. Structural processes at slow-spreading ridges[J]. Science, 1992, 257(5070): 627-634. PMID