Advanced Search
Article Contents

Improvement and Evaluation of the Latest Version of WRF-Lake at a Deep Riverine Reservoir


doi: 10.1007/s00376-022-2180-5

  • The WRF-lake vertically one-dimensional (1D) water temperature model, as a submodule of the Weather Research and Forecasting (WRF) system, is being widely used to investigate water–atmosphere interactions. But previous applications revealed that it cannot accurately simulate the water temperature in a deep riverine reservoir during a large flow rate period, and whether it can produce sufficiently accurate heat flux through the water surface of deep riverine reservoirs remains uncertain. In this study, the WRF-lake model was improved for applications in large, deep riverine reservoirs by parametric scheme optimization, and the accuracy of heat flux calculation was evaluated compared with the results of a better physically based model, the Delft3D-Flow, which was previously applied to different kinds of reservoirs successfully. The results show: (1) The latest version of WRF-lake can describe the surface water temperature to some extent but performs poorly in the large flow period. We revised WRF-lake by modifying the vertical thermal diffusivity, and then, the water temperature simulation in the large flow period was improved significantly. (2) The latest version of WRF-lake overestimates the reservoir–atmosphere heat exchange throughout the year, mainly because of underestimating the downward energy transfer in the reservoir, resulting in more heat remaining at the surface and returning to the atmosphere. The modification of vertical thermal diffusivity can improve the surface heat flux calculation significantly. (3) The longitudinal temperature variation and the temperature difference between inflow and outflow, which cannot be considered in the 1D WRF-lake, can also affect the water surface heat flux.
    摘要: WRF-lake是简化的垂向一维水温模型,作为WRF气象模式的子模块,广泛应用于水体与大气间交互过程的模拟预测。已有研究表明,WRF-lake不能准确模拟河道型大深水库在大流量时的水温;WRF-lake能否准确计算河道型大深水库与大气间的热通量,目前也不清楚。本研究首先对WRF-lake模型的参数化方案进行改进,提高其在河道型大深水库中的水温模拟精度,并将其与物理过程更完整且已成功应用于各类水库的Delft3D-Flow模型进行对比,评估了WRF-lake对河道型大深水库表面热通量的模拟准确性。结果表明:(1)虽然最新版本的WRF-lake模型能够在一定程度上准确模拟河道型大深水库的表面水温,但在大流量时段表现较差。本研究提出的修正垂向热扩散系数方案,可显著改善大流量时段的水温模拟结果。(2)最新版本的WRF-lake模型全年都高估了水库表面的热交换量,这主要由于低估了水库中能量的向下传输,导致更多热量滞留在水体表面并返回到大气中。本研究对垂向热扩散系数的改进,提高了表面热通量的计算精度。(3)作为简化的垂向一维模型,WRF-lake不能反映水温的纵向变化和出入流温差,这两个因素也对水体表面热通量具有一定影响。
  • 加载中
  • Figure 1.  Simplification of a narrow, long, and deep riverine reservoir represented as a vertically 1D barrel-shaped model in the WRF-lake model.

    Figure 2.  The Lancang-Mekong River basin and the locations of cascade reservoirs on the middle-lower reach.

    Figure 3.  Monthly averaged inflow-outflow rates and inflow temperature of the Nuozhadu Reservoir in 2015; Daily averaged meteorological data of the Nuozhadu Reservoir in 2015.

    Figure 4.  The morphologic grid of the Nuozhadu Reservoir.

    Figure 5.  Simulated surface water temperature in 2015, including the results of the WRF-lake model version by Wang et al. (2019) (blue solid line), the modified WRF-lake model used in this paper (blue dotted line), and the Delft3D-Flow simulation (black solid line), with the observed data (red dots) as reference.

    Figure 6.  (a)Vertical water temperature profiles of the WRF-lake model version by Wang et al. (2019) (blue lines) and the modified WRF-lake model used in this paper (black lines). The dots represent observed data. (b) Diffusion coefficient profiles of the WRF-lake model version by Wang et al. (2019) (blue lines) and the modified WRF-lake model used in this paper (black lines).

    Figure 7.  Simulated water temperature distributions in April, July, September, and December of 2015 by Delft3D-Flow, representing spring, summer, autumn, and winter, respectively.

    Figure 8.  Different kinds of heat flux per unit area through the water surface, including net longwave radiation, latent heat, and sensible heat calculated by the WRF-lake model version by Wang et al. (2019) (red lines), the modified WRF-lake model used in this paper (blue lines), and the Delft3D-Flow simulation (black lines). Positive values and negative values represent the reservoir releasing energy to the atmosphere and absorbing energy from the atmosphere

    Figure 6.  (Continued)

    Table 1.  The RMSE and MAE of the WRF-lake model version by Wang et al. (2019), the modified WRF-lake model used in this paper, and the Delft3D-Flow model used in surface water temperature and vertical temperature profile simulation near the dam.

    WRF-lake [by Wang et al. (2019)]WRF-lake (modified)Delft3D-Flow
    Surface water temperatureRMSE (°C)1.501.081.19
    MAE (°C)0.350.150.10
    Vertical water temperature profileRMSE (°C)1.131.011.10
    MAE (°C)0.940.700.60
    DownLoad: CSV

    Table 2.  Annual average values and average values of the large flow rate period, i.e., from May to August, for different kinds of heat fluxes per unit area through the water surface, including net longwave radiation, latent heat, and sensible heat calculated by the WRF-lake model version by Wang et al. (2019), the modified WRF-lake model used in this paper, and the Delft3D-Flow simulation. Positive values represent the reservoir releasing energy to the atmosphere.

    Net Longwave Radiation (W m–2)Sensible Heat (W m–2)Latent Heat (W m–2)
    Whole yearLarge flow rate periodWhole yearLarge flow rate periodWhole yearLarge flow rate period
    WRF-lake [by Wang et al. (2019)]112.3107.717.414.382.586.2
    WRF-lake (modified)100.382.211.55.660.946.6
    Delft3D-Flow88.484.09.05.950.348.2
    DownLoad: CSV
  • Adrian, R., and Coauthors, 2009: Lakes as sentinels of climate change. Limnology and Oceanography, 54, 2283−2297, https://doi.org/10.4319/lo.2009.54.6_part_2.2283.
    Bates, G. T., F. Giorgi, and S. W. Hostetler, 1993: Toward the simulation of the effects of the great lakes on regional climate. Mon. Wea. Rev., 121, 1373−1387, https://doi.org/10.1175/1520-0493(1993)121<1373:TTSOTE>2.0.CO;2.
    Bonan, G. B., 1995: Sensitivity of a GCM simulation to inclusion of inland water surfaces. J. Climate, 8, 2691−2704, https://doi.org/10.1175/1520-0442(1995)008<2691:SOAGST>2.0.CO;2.
    Campbell, I. C., 2007: Perceptions, data, and river management: Lessons from the Mekong River. Water Resour. Res., 43, W02407, https://doi.org/10.1029/2006WR005130.
    Carrivick, J. L., L. E. Brown, D. M. Hannah, and A. G. D. Turner, 2012: Numerical modelling of spatio-temporal thermal heterogeneity in a complex river system. J. Hydrol., 414−415, 491−502,
    Carron, J. C., and H. Rajaram, 2001: Impact of variable reservoir releases on management of downstream water temperatures. Water Resour. Res, 37, 1733−1743, https://doi.org/10.1029/2000WR900390.
    Chanudet, V., V. Fabre, and T. V. D. Kaaij, 2012: Application of a three-dimensional hydrodynamic model to the Nam Theun 2 Reservoir (Lao PDR). Journal of Great Lakes Research, 38, 260−269, https://doi.org/10.1016/j.jglr.2012.01.008.
    Chow, V. T., 1959: Open-Channel Hydraulics. McGraw-Hill.
    Cipagauta, C., B. Mendoza, and J. Zavala-Hidalgo, 2014: Sensitivity of the surface temperature to changes in total solar irradiance calculated with the WRF model. Geofisica Internacional, 53, 153−162, https://doi.org/10.1016/S0016-7169(14)71497-7.
    Davis, C., and Coauthors, 2008: Prediction of landfalling hurricanes with the advanced hurricane WRF model. Mon. Wea. Rev., 136, 1990−2005, https://doi.org/10.1175/2007MWR2085.1.
    Deng, B., S. D. Liu, W. Xiao, W. Wang, J. M. Jin, and X. Lee, 2013: Evaluation of the CLM4 lake model at a large and shallow freshwater lake. Journal of Hydrometeorology, 14, 636−649, https://doi.org/10.1175/jhm-d-12-067.1.
    Deltares, 2003: Delft3D-FLOW: Simulation of Multi-Dimensional Hydrodynamic Flows and Transport Phenomena, Including Sediments.
    Ellis, C. R., H. G. Stefan, and R. C. Gu, 1991: Water temperature dynamics and heat transfer beneath the ice cover of a lake. Limnology and Oceanography, 36, 324−334, https://doi.org/10.4319/lo.1991.36.2.0324.
    Elzawahry, E. A., 1985: Advection, diffusion and settling in the coastal boundary layer of Lake Erie. PhD dissertation, McMaster University.
    Fan, H., D. M. He, and H. L. Wang, 2015: Environmental consequences of damming the mainstream Lancang-Mekong River: A review. Earth-Science Reviews, 146, 77−91, https://doi.org/10.1016/j.earscirev.2015.03.007.
    Fang, N., K. Yang, Z. La, Y. Y. Chen, J. B. Wang, and L. P. Zhu, 2017: Research on the application of WRF-Lake modeling at Nam Co Lake on the Qinghai-Tibetan Plateau. Plateau Meteorology, 36, 610−618, https://doi.org/10.7522/j.issn.1000-0534.2016.00038. (in Chinese with English abstract
    Gerken, T., W. Babel, F. L. Sun, M. Herzog, Y. M. Ma, T. Foken, and H. F. Graf, 2013: Uncertainty in atmospheric profiles and its impact on modeled convection development at Nam Co Lake, Tibetan Plateau. J. Geophys. Res.: Atmos., 118, 12 317−12 331,
    Gillett, R. M., and P. V. Tobias, 2002: Human growth in southern Zambia: A first study of Tonga children predating the Kariba dam (1957-1958). American Journal of Human Biology, 14, 50−60, https://doi.org/10.1002/ajhb.10019.
    Gorham, E., and F. M. Boyce, 1989: Influence of lake surface area and depth upon thermal stratification and the depth of the summer thermocline. Journal of Great Lakes Research, 15, 233−245, https://doi.org/10.1016/S0380-1330(89)71479-9.
    Gu, H. P., Z. G. Ma, and M. X. Li, 2016: Effect of a large and very shallow lake on local summer precipitation over the Lake Taihu basin in China. J. Geophys. Res.: Atmos., 121, 8832−8848, https://doi.org/10.1002/2015JD024098.
    Gu, H. P., J. M. Jin, Y. H. Wu, M. B. Ek, and Z. M. Subin, 2015: Calibration and validation of lake surface temperature simulations with the coupled WRF-Lake model. Climate Change, 129, 471−483, https://doi.org/10.1007/s10584-013-0978-y.
    Gu, J. F., Q. N. Xiao, Y. H. Kuo, D. M. Barker, J. S. Xue, and X. X. Ma, 2005: Assimilation and simulation of typhoon Rusa (2002) using the WRF system. Adv. Atmos. Sci., 22, 415−427, https://doi.org/10.1007/BF02918755.
    Guo, M. Y., Q. L. Zhuang, H. X. Yao, M. Golub, L. R. Leung, D. Pierson, and Z. L. Tan, 2021: Validation and sensitivity analysis of a 1-D lake model across global lakes. J. Geophys. Res.: Atmos., 126, E2020JD033417, https://doi.org/10.1029/2020jd033417.
    Guo, S. B., F. S. Wang, D. J. Zhu, G. H. Ni, and Y. C. Chen, 2022: Evaluation of the WRF-lake model in the large dimictic reservoir: Comparisons with field data and another water temperature model. Journal of Hydrometeorology, 23, 1227−1244, https://doi.org/10.1175/JHM-D-21-0220.1.
    Henderson-Sellers, B., 1985: New formulation of eddy diffusion thermocline models. Applied Mathematical Modelling, 9, 441−446, https://doi.org/10.1016/0307-904X(85)90110-6.
    Henderson-Sellers, B., M. J. Mccormick, and D. Scavia, 1983: A comparison of the formulation for eddy diffusion in two one-dimensional stratification models. Applied Mathematical Modelling, 7, 212−215, https://doi.org/10.1016/0307-904X(83)90010-0.
    Hostetler, S. W., G. T. Bates, and F. Giorgi, 1993: Interactive coupling of a lake thermal model with a regional climate model. J. Geophys. Res.: Atmos., 98, 5045−5057, https://doi.org/10.1029/92JD02843.
    Hostetler, S. W., F. Giorgi, G. T. Bates, and P. J. Bartlein, 1994: Lake-atmosphere feedbacks associated with paleolakes Bonneville and Lahontan. Science, 263(5147), 665−668, https://doi.org/10.1126/science.263.5147.665.
    Huang, A. N., and Coauthors, 2019: Evaluating and improving the performance of three 1-D lake models in a large deep lake of the central Tibetan Plateau. J. Geophys. Res.: Atmos., 124, 3143−3167, https://doi.org/10.1029/2018JD029610.
    Imberger, J., 1985: The diurnal mixed layer. Limnology and Oceanography, 30, 737−770, https://doi.org/10.4319/lo.1985.30.4.0737.
    Jiang, B., F. S. Wang, and G. H. Ni, 2018: Heating impact of a tropical reservoir on downstream water temperature: A case study of the Jinghong Dam on the Lancang River. Water, 10, 951, https://doi.org/10.3390/w10070951.
    Li, Z. S., D. J. Zhu, Y. C. Chen, X. Fang, Z. W. Liu, and W. Ma, 2016: Simulating and understanding effects of water level fluctuations on thermal regimes in Miyun reservoir. Hydrological Sciences Journal, 61, 952−969, https://doi.org/10.1080/02626667.2014.983517.
    Ma, D., T. Y. Wang, C. Gao, S. L. Pan, Z. L. Sun, and Y. P. Xu, 2018: Potential evapotranspiration changes in Lancang River basin and Yarlung Zangbo River basin, southwest China. Hydrological Sciences Journal, 63, 1653−1668, https://doi.org/10.1080/02626667.2018.1524147.
    Mackay, M. D., and Coauthors, 2009: Modeling lakes and reservoirs in the climate system. Limnology and Oceanography, 54, 2315−2329, https://doi.org/10.4319/lo.2009.54.6_part_2.2315.
    Mallard, M. S., C. G. Nolte, T. L. Spero, O. R. Bullock, K. Alapaty, J. A. Herwehe, J. Gula, and J. H. Bowden, 2015: Technical challenges and solutions in representing lakes when using WRF in downscaling applications. Geoscientific Model Development, 8, 1085−1096, https://doi.org/10.5194/gmd-8-1085-2015.
    Martynov, A., L. Sushama, and R. Laprise, 2010: Simulation of temperate freezing lakes by one-dimensional lake models: Performance assessment for interactive coupling with regional climate models. Boreal Environment Research, 15, 143−164.
    Milly, P. C. D., J. Betancourt, M. Falkenmark, R. M. Hirsch, Z. W. Kundzewicz, D. P. Lettenmaier, and R. J. Stouffer, 2008: Stationarity is dead: Whither water management? Science, 319, 573−574, https://doi.org/10.1126/science.1151915.
    Nowlin, W. H., J. M. Davies, R. N. Nordin, and A. Mazumder, 2004: Effects of water level fluctuation and short-term climate variation on thermal and stratification regimes of a British Columbia reservoir and lake. Lake and Reservoir Management, 20, 91−109, https://doi.org/10.1080/07438140409354354.
    Owens, E. M., 1998: Thermal and heat transfer characteristics of Cannonsville Reservoir. Lake and Reservoir Management, 14, 152−161, https://doi.org/10.1080/07438149809354327.
    Skamarock, W. C., and Coauthors, 2008: A description of the advanced research WRF version 3. NCAR Tech. Note NCAR/TN-475 + STR,
    Stepanenko, V. M., and Coauthors, 2013: A one-dimensional model intercomparison study of thermal regime of a shallow, turbid midlatitude lake. Geoscientific Model Development, 6, 1337−1352, https://doi.org/10.5194/gmd-6-1337-2013.
    Straškraba, M., J. G. Tundisi, and A. Duncan, 1993: State-of-the-art of reservoir limnology and water quality management. Comparative Reservoir Limnology and Water Quality Management, Springer, 213−288,
    Strzepek, K. M., G. W. Yohe, R. S. J. Tol, and M. W. Rosegrant, 2008: The value of the high Aswan dam to the Egyptian economy. Ecological Economics, 66, 117−126, https://doi.org/10.1016/j.ecolecon.2007.08.019.
    Subin, Z. M., W. J. Riley, and D. Mironov, 2012: An improved lake model for climate simulations: Model structure, evaluation, and sensitivity analyses in CESM1. Journal of Advances in Modeling Earth Systems, 4, M02001, https://doi.org/10.1029/2011MS000072.
    Swayne, D., D. Lam, M. Mackay, W. Rouse, and W. Schertzer, 2005: Assessment of the interaction between the Canadian Regional Climate Model and lake thermal-hydrodynamic models. Environmental Modelling & Software, 20, 1505−1513, https://doi.org/10.1016/j.envsoft.2004.08.015.
    Sweers, H. E., 1970: Vertical diffusivity coefficient in a thermocline. Limnology and Oceanography, 15, 273−280, https://doi.org/10.4319/lo.1970.15.2.0273.
    Tsanis, I. K., and J. Wu, 2000: Application and verification of a three-dimensional hydrodynamic model to Hamilton Harbour, Canada. Global Nest Journal, 2, 77−89.
    Van der Kaaij, T., D. Roelvink, and C. Kuijper, 2004: Morphological modelling of the western Scheldt. Intermediate report phase II: Calibration of the morphological model. Alkyon/WL | Delft Hydraulics: Delft.
    Wang, F. S., G. H. Ni, W. J. Riley, J. Y. Tang, D. J. Zhu, and T. Sun, 2019: Evaluation of the WRF lake module (v1.0) and its improvements at a deep reservoir. Geoscientific Model Development, 12, 2119−2138, https://doi.org/10.5194/gmd-12-2119-2019.
    Xiao, C. L., B. M. Lofgren, J. Wang, and P. Y. Chu, 2016: Improving the lake scheme within a coupled WRF-lake model in the Laurentian Great Lakes. Journal of Advances in Modeling Earth Systems, 8, 1969−1985, https://doi.org/10.1002/2016MS000717.
    Xie, Q. K., Z. W. Liu, X. Fang, Y. C. Chen, C. Li, and S. Macintyre, 2017: Understanding the temperature variations and thermal structure of a subtropical deep river-run reservoir before and after impoundment. Water, 9, 603, https://doi.org/10.3390/w9080603.
    Xing, Z. K., D. A. Fong, K. M. Tan, E. Y. M. Lo, and S. G. Monismith, 2012: Water and heat budgets of a shallow tropical reservoir. Water Resour. Res., 48, W06532, https://doi.org/10.1029/2011WR011314.
    Xu, L. J., H. Z. Liu, Q. Du, and L. Wang, 2016: Evaluation of the WRF-lake model over a highland freshwater lake in southwest China. J. Geophys. Res.: Atmos., 121, 13 989−14 005,
    Yao, Y., L. Ruan, H. Li, W. J. Zhou, D. Yang, and J. H. Yu, 2013: Changes of meteorological parameters and lightning current during water impounded in Three Gorges area. Atmospheric Research, 134, 150−160, https://doi.org/10.1016/j.atmosres.2013.06.004.
    Zhu, K. F., and M. Xue, 2016: Evaluation of WRF-based convection-permitting multi-physics ensemble forecasts over China for an extreme rainfall event on 21 July 2012 in Beijing. Adv. Atmos. Sci., 33, 1240−1258, https://doi.org/10.1007/s00376-016-6202-z.
    Zolfaghari, K., C. R. Duguay, and H. K. Pour, 2017: Satellite-derived light extinction coefficient and its impact on thermal structure simulations in a 1-D lake model. Hydrology and Earth System Sciences, 21, 377−391, https://doi.org/10.5194/hess-21-377-2017.
  • [1] ZENG Xiaodong, WANG Aihui, ZENG Qingcun, Robert E. DICKINSON, Xubin ZENG, Samuel S. P. SHEN, 2006: Intermediately Complex Models for the Hydrological Interactions in the Atmosphere-Vegetation-Soil System, ADVANCES IN ATMOSPHERIC SCIENCES, 23, 127-140.  doi: 10.1007/s00376-006-0013-6
    [2] Bao Ning, Zhang Xuehong, 1991: Effect of Ocean Thermal Diffusivity on Global Warming Induced by Increasing Atmospheric CO2, ADVANCES IN ATMOSPHERIC SCIENCES, 8, 421-430.  doi: 10.1007/BF02919265
    [3] JIN Ling, Fanyou KONG, LEI Hengchi*, and HU Zhaoxia, 2014: A Methodological Study on Using Weather Research and Forecasting (WRF) Model Outputs to Drive a One-Dimensional Cloud Model, ADVANCES IN ATMOSPHERIC SCIENCES, 31, 230-240.  doi: 10.1007/s00376-013-2257-2
    [4] Hu Yinqiao, Su Congxian, Zhang Yongfeng, 1988: RESEARCH ON THE MICROCLIMATE CHARACTERISTICS AND COLD ISLAND EFFECT OVER A RESERVOIR IN THE HEXI REGION, ADVANCES IN ATMOSPHERIC SCIENCES, 5, 117-126.  doi: 10.1007/BF02657351
    [5] Wang Shaowu, 1984: THE RHYTHM IN THE ATMOSPHERE AND OCEANS IN APPLICATION TO LONG-RANGE WEATHER FORECASTING, ADVANCES IN ATMOSPHERIC SCIENCES, 1, 7-29.  doi: 10.1007/BF03187612
    [6] Yong. L. McHall, 1991: Planetary Stationary Waves in the Atmosphere Part II: Thermal Stationary Waves, ADVANCES IN ATMOSPHERIC SCIENCES, 8, 225-236.  doi: 10.1007/BF02658096
    [7] Li Xianlang, 1988: NONLINEAR RESONANCE INTERACTIONS AND INDEX CYCLES IN THE ATMOSPHERE, ADVANCES IN ATMOSPHERIC SCIENCES, 5, 253-264.  doi: 10.1007/BF02656750
    [8] LIU Huizhi, LIANG Bin, ZHU Fengrong, ZHANG Boyin, SANG Jianguo, 2004: Water-Tank Experiment on the Thermal Circulation Induced by the Bottom Heating in an Asymmetric Valley, ADVANCES IN ATMOSPHERIC SCIENCES, 21, 536-546.  doi: 10.1007/BF02915721
    [9] Shaofeng LIU, Michael HINTZ, Xiaolong LI, 2016: Evaluation of Atmosphere-Land Interactions in an LES from the Perspective of Heterogeneity Propagation, ADVANCES IN ATMOSPHERIC SCIENCES, 33, 571-578.  doi: 10.1007/s00376-015-5212-6
    [10] XU Xingkui, Jason K. LEVY, 2011: The Impact of Agricultural Practices in China on Land-Atmosphere Interactions, ADVANCES IN ATMOSPHERIC SCIENCES, 28, 821-831.  doi: 10.1007/s00376-010-0007-2
    [11] Zhu Xun, 1987: ON GRAVITY WAVE-MEAN FLOW INTERACTIONS IN A THREE DIMENSIONAL STRATIFIED ATMOSPHERE, ADVANCES IN ATMOSPHERIC SCIENCES, 4, 287-299.  doi: 10.1007/BF02663599
    [12] Xiaojuan LIU, Guangjin TIAN, Jinming FENG, Bingran MA, Jun WANG, Lingqiang KONG, 2018: Modeling the Warming Impact of Urban Land Expansion on Hot Weather Using the Weather Research and Forecasting Model: A Case Study of Beijing, China, ADVANCES IN ATMOSPHERIC SCIENCES, 35, 723-736.  doi: 10.1007/s00376-017-7137-8
    [13] Peter C. Chu, Chen Yuchun, Lu Shihua, 2001: Evaluation of Haney-Type Surface Thermal Boundary Conditions Using a Coupled Atmosphere and Ocean Model, ADVANCES IN ATMOSPHERIC SCIENCES, 18, 355-375.  doi: 10.1007/BF02919315
    [14] Qin XU, Wei GU, GAO Shouting, 2005: Nonlinear Oscillations of Semigeostrophic Eady Waves in the Presence of Diffusivity, ADVANCES IN ATMOSPHERIC SCIENCES, 22, 49-57.  doi: 10.1007/BF02930869
    [15] DONG Danpeng, ZHOU Weidong, YANG Yang, DU Yan, 2010: On Outflow Passages in the South China Sea, ADVANCES IN ATMOSPHERIC SCIENCES, 27, 60-68.  doi: 10.1007/s00376-009-8050-6
    [16] LUO Yunfeng, ZHENG Wenjing, ZHOU Xiaogang, 2004: Research Project Entitled "The Dynamics and Physical Processes in The Weather and Climate System" --Part Ⅰ: A Brief Introduction, ADVANCES IN ATMOSPHERIC SCIENCES, 21, 824-829.  doi: 10.1007/BF02916378
    [17] Rui WANG, Yiting ZHU, Fengxue QIAO, Xin-Zhong LIANG, Han ZHANG, Yang DING, 2021: High-resolution Simulation of an Extreme Heavy Rainfall Event in Shanghai Using the Weather Research and Forecasting Model: Sensitivity to Planetary Boundary Layer Parameterization, ADVANCES IN ATMOSPHERIC SCIENCES, 38, 98-115.  doi: 10.1007/s00376-020-9255-y
    [18] Rajabu J. MANGARA, Zhenhai GUO, Shuanglin LI, 2019: Performance of the Wind Farm Parameterization Scheme Coupled with the Weather Research and Forecasting Model under Multiple Resolution Regimes for Simulating an Onshore Wind Farm, ADVANCES IN ATMOSPHERIC SCIENCES, 36, 119-132.  doi: 10.1007/s00376-018-8028-3
    [19] SUN Shufen, YAN Jinfeng, XIA Nan, SUN Changhai, 2007: Development of a Model for Water and Heat Exchange Between the Atmosphere and a Water Body, ADVANCES IN ATMOSPHERIC SCIENCES, 24, 927-938.  doi: 10.1007/s00376-007-0927-7
    [20] M.Lal, 1994: Water Resources of the South Asian Region in a Warmer Atmosphere, ADVANCES IN ATMOSPHERIC SCIENCES, 11, 239-246.  doi: 10.1007/BF02666550

Get Citation+

Export:  

Share Article

Manuscript History

Manuscript received: 26 June 2022
Manuscript revised: 31 August 2022
Manuscript accepted: 13 September 2022
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Improvement and Evaluation of the Latest Version of WRF-Lake at a Deep Riverine Reservoir

    Corresponding author: Dejun ZHU, zhudejun@tsinghua.edu.cn
  • 1. State Key Laboratory of Hydroscience and Engineering, Tsinghua University, Beijing 100084, China
  • 2. Southwest Petroleum University, Chengdu 610599, Sichuan, China

Abstract: The WRF-lake vertically one-dimensional (1D) water temperature model, as a submodule of the Weather Research and Forecasting (WRF) system, is being widely used to investigate water–atmosphere interactions. But previous applications revealed that it cannot accurately simulate the water temperature in a deep riverine reservoir during a large flow rate period, and whether it can produce sufficiently accurate heat flux through the water surface of deep riverine reservoirs remains uncertain. In this study, the WRF-lake model was improved for applications in large, deep riverine reservoirs by parametric scheme optimization, and the accuracy of heat flux calculation was evaluated compared with the results of a better physically based model, the Delft3D-Flow, which was previously applied to different kinds of reservoirs successfully. The results show: (1) The latest version of WRF-lake can describe the surface water temperature to some extent but performs poorly in the large flow period. We revised WRF-lake by modifying the vertical thermal diffusivity, and then, the water temperature simulation in the large flow period was improved significantly. (2) The latest version of WRF-lake overestimates the reservoir–atmosphere heat exchange throughout the year, mainly because of underestimating the downward energy transfer in the reservoir, resulting in more heat remaining at the surface and returning to the atmosphere. The modification of vertical thermal diffusivity can improve the surface heat flux calculation significantly. (3) The longitudinal temperature variation and the temperature difference between inflow and outflow, which cannot be considered in the 1D WRF-lake, can also affect the water surface heat flux.

摘要: WRF-lake是简化的垂向一维水温模型,作为WRF气象模式的子模块,广泛应用于水体与大气间交互过程的模拟预测。已有研究表明,WRF-lake不能准确模拟河道型大深水库在大流量时的水温;WRF-lake能否准确计算河道型大深水库与大气间的热通量,目前也不清楚。本研究首先对WRF-lake模型的参数化方案进行改进,提高其在河道型大深水库中的水温模拟精度,并将其与物理过程更完整且已成功应用于各类水库的Delft3D-Flow模型进行对比,评估了WRF-lake对河道型大深水库表面热通量的模拟准确性。结果表明:(1)虽然最新版本的WRF-lake模型能够在一定程度上准确模拟河道型大深水库的表面水温,但在大流量时段表现较差。本研究提出的修正垂向热扩散系数方案,可显著改善大流量时段的水温模拟结果。(2)最新版本的WRF-lake模型全年都高估了水库表面的热交换量,这主要由于低估了水库中能量的向下传输,导致更多热量滞留在水体表面并返回到大气中。本研究对垂向热扩散系数的改进,提高了表面热通量的计算精度。(3)作为简化的垂向一维模型,WRF-lake不能反映水温的纵向变化和出入流温差,这两个因素也对水体表面热通量具有一定影响。

    • Inland waters like lakes, rivers, and reservoirs play important roles in local and regional climate. Compared with land covers, water surfaces have different physical properties, such as lower albedo, lower surface roughness, higher heat capacity, etc. The mixing mechanism in water is complicated, including turbulent diffusion and convective mixing, etc., resulting in complicated heat circulating characteristics. Through affecting the fluxes of heat and moisture, water can modulate local and regional atmospheric boundary layer conditions. For example, in summer, water always absorbs energy from the atmosphere and decreases air temperature. In winter, the energy collected in summer is released to the atmosphere, consequently increasing the air temperature (Bates et al., 1993; Hostetler et al., 1994; Bonan, 1995; Swayne et al., 2005; Milly et al., 2008; Adrian et al., 2009; Martynov et al., 2010; Gerken et al., 2013; Zolfaghari et al., 2017).

      The Weather Research and Forecasting (WRF) system is a mesoscale numerical atmospheric model that has been widely used in recent years to analyze large-scale hydrology characteristics, examine water–atmosphere interactions, and forecast weather (Gu et al., 2005, 2015; Davis et al., 2008; Skamarock et al., 2008; Cipagauta et al., 2014; Zhu and Xue, 2016; Fang et al., 2017). A vertically one-dimensional (1D) water temperature model, WRF-lake, has been coupled as a submodule to the WRF system (since version 3.6) by Gu et al. (2015) to describe the thermal regimes in water and in turn better calculate the heat flux between water and atmosphere (Gu et al., 2015, 2016; Mallard et al., 2015; Xiao et al., 2016). The WRF-lake model was developed from the concept of Henderson-Sellers’s eddy diffusion thermocline model (Henderson-Sellers et al., 1983; Henderson-Sellers, 1985), simplifying the lake to a 1D barrel-shaped model and dividing the water body into several layers vertically (Hostetler et al., 1993; Hostetler et al., 1994). The thermal distribution was simulated by solving the vertically 1D thermal diffusion equation (Bonan, 1995; Subin et al., 2012). Sediment, ice, and snow layers were added to the model by Subin et al. (2012) based on the parameterization scheme of the CLM4.LISSS model. After a series of improvements, including modifying the lake surface albedo, the vertical eddy diffusivity, the extinction of solar radiation, the surface roughness, etc., it has been proven that the latest version of WRF-lake can be applied to many kinds of natural lakes (Gu et al., 2015; Xiao et al., 2016; Xu et al., 2016; Fang et al., 2017; Huang et al., 2019; Guo et al., 2022).

      However, there have been few applications of the WRF-lake model in manmade reservoirs, especially large, deep riverine reservoirs, which compared with natural lakes, have more complex thermal characteristics (Carron and Rajaram, 2001; Li et al., 2016). Because of larger inflow/outflow rates, more frequent changes of water level, and more complicated topography, riverine reservoirs undergo significantly larger variations of thermal structure, which in turn exert significant impacts on regional water–atmosphere interactions (Owens, 1998; Nowlin et al., 2004; Xing et al., 2012; Deng et al., 2013; Stepanenko et al., 2013; Yao et al., 2013). Wang et al. (2019) found that the original WRF-lake model could not predict a satisfactory water temperature profile near the dam of a deep riverine reservoir and improved this through optimizing the parameterization schemes. However, the results of Wang et al. (2019) still exhibited obvious bias during periods of large flow. This may be due to a failure to take into account the effect of inflow-outflow. Guo et al. (2022) applied the WRF-lake model in a relatively small and shallow reservoir with relatively small flow rate, i.e., the Miyun Reservoir in Northern China, and tested the effect of inflow-outflow. In the Miyun Reservoir, since the flow rate is small, i.e., less than 50 m3 s–1, the effect of inflow-outflow on the surface water temperature is quite small. How the WRF-lake model performs in large, deep, and riverine reservoirs like the Nuozhadu Reservoir, which has an annual average flow rate of about 1500 m3 s–1, still needs to be explored.

      Moreover, in a riverine reservoir, there may be non-negligible longitudinal variation of water temperature (Fig. 1). Although 1D models apparently seem oversimplified compared with multidimensional models accounting for more accurate spatiotemporal dynamics, they have lower data requirements, higher computational efficiency, and lower complexity for coupling, which are meaningful for studies of regional scales and weather forecasting timeliness (Mackay et al., 2009; Gu et al., 2015; Wang et al., 2019; Guo et al., 2021). The aforementioned Miyun Reservoir has a lake-like plan form, and whether the heat flux calculated by the 1D WRF-lake model in a riverine reservoir is reasonable has yet to be studied.

      Figure 1.  Simplification of a narrow, long, and deep riverine reservoir represented as a vertically 1D barrel-shaped model in the WRF-lake model.

      The main aims of this study are as follows:

      (1) To improve the performance of the WRF-lake water temperature model in large, deep riverine reservoirs during large flow rate periods by parametric scheme optimization;

      (2) To evaluate the accuracy of heat flux calculated by WRF-lake and find out whether the 1D model is applicable to studying interactions between large, deep riverine reservoirs and atmosphere.

      To achieve the two objectives, we applied the WRF-lake model to a deep riverine reservoir, namely the Nuozhadu Reservoir in Yunnan province of China, which was also studied by Wang et al. (2019). First, we modified the parameterization of the vertical thermal diffusivity to make the results agree better with the observed data, especially in periods of large flow rate. Then, we compared the simulated water–atmosphere heat flux by WRF-lake with the results of a better, physically based three-dimensional (3D) model, the Delft3D-Flow, which had previously been applied to the Nuozhadu and other large reservoirs successfully (Jiang et al., 2018; Wang et al., 2019). The Delft3D-Flow model, which is more complicated but more accurate, can sufficiently illustrate the variation of water temperature and surface heat flux, and the results can serve as a reference to evaluate the performance of the 1D WRF-lake model. The simulation of WRF-lake was done offline in order to ensure the meteorologically driven data was consistent between both WRF-lake and Delft3D-Flow. The surface heat fluxes in a unit area calculated by Delft3D-Flow and WRF-lake were analyzed and compared.

      The Nuozhadu Reservoir, as a large riverine reservoir with a large flow rate and significant vertical thermal stratification, is typical. There are many reservoirs around the world with similar characteristics, such as the Aswan Reservoir in Egypt (Strzepek et al., 2008), the Kariba Reservoir in Zambezi (Gillett and Tobias, 2002), and the Xiluodu Reservoir in Sichuan China (Xie et al., 2017), which are all famous and widely studied. Therefore, the case for the Nuozhadu Reservoir, as a representative study, can serve as a reference for the studies in other large riverine reservoirs.

      The rest of this paper is organized as follows: The model description and an explanation of its improvement are given in section 2. The study area, datasets, and model settings are introduced in section 3. The results are presented and discussed in section 4. Finally, the summary and conclusions are given in section 5.

    2.   Model description and improvement
    • The WRF-lake model simplifies a lake or a reservoir to a vertically 1D barrel-shaped model and divides the water body into layers with varying vertical discretization. Thermal processes through the water surface considered in WRF-lake include longwave radiation, shortwave radiation, sensible heat, and latent heat. Snow and ice on top of the water column that can influence the simulation of downward heat transfer are considered. Soil and bedrock at the bottom are also considered.

      The equation of heat flux through the water surface in WRF-lake, which is the same as in the Delft3D-Flow model, is as follows:

      where $ \beta $ is surface absorption fraction of the net shortwave radiation S (W m–2), L is surface absorption of net longwave radiation (W m–2), H and E are sensible heat (W m–2) and latent heat (W m–2), respectively, and G is downward heat transfer (W m–2). These terms are calculated by Eqs. (2)–(5) as follows:

      In Eq. (2), ε is lake emissivity, σ is the Stefan-Boltzmann constant (W m–2 K–4), ${T}_{{\rm{g}}}$ is surface water temperature (°C), and ${L}_{{\rm{atm}}}$ is downward longwave radiation (W m–2). In Eq. (3), ${\rho }_{{\rm{atm}}}$ is moist atmospheric density (kg m–3), cp is specific heat of air (J kg–1 K–1), ${\theta }_{{\rm{atm}}}$ is potential temperature at atmospheric reference (K), and rah is aerodynamic resistance to latent heat (m–1 s–1). In Eq. (4), λ is latent heat of vaporization and sublimation depending on the surface water temperature (J kg–1 s–2), ${q}_{{\rm{g}}}$ and ${q}_{{\rm{atm}}}$ are saturated specific humidity at the surface (kg–1 kg–1) and specific humidity at the reference height (kg–1 kg–1), respectively, and raw is aerodynamic resistance to sensible heat (m–1 s–1). And in Eq. (5), k is the thermal conductivity of the top layer (W m–1 K–1), ${T}_{{\rm{T}}}$ is the water temperature of the top layer (K), and ${\Delta z}_{{\rm{T}}}$ is the thickness of the top layer (m).

      Vertical heat exchanges between different water layers are calculated by solving the 1D unsteady diffusion equation as follows (Gu et al., 2015):

      where K is the vertical eddy diffusion coefficient, used to describe the vertical heat exchange between layers, T is water temperature of the corresponding layer (°C), t is time (s), ${c}_{{\rm{liq}}}$ is the volumetric heat capacity of water (J kg–1 °C–1), z is the depth under the water surface (m), and $ \varnothing $ is the solar radiation heat source term (W m–2) (Subin et al., 2012). And K is made up of two parts:

      where ${K}_{{\rm{m}}}$ is the molecular diffusion coefficient, to be 1.433 × 10−7 m2 s–1, and ${K}_{{\rm{e}}}$ is the wind-driven eddy diffusion coefficient (Henderson-Sellers et al., 1983; Henderson-Sellers, 1985), calculated as follows:

      where k is the von Karman constant, w is wind speed above the water surface (m s–1), P0 is the neutral turbulent Prandtl number, $ {R}_{i} $ is the gradient Richardson number, ${T}_{{\rm{g}}}$ is surface water temperature (°C), ${T}_{{\rm{f}}}$ is freezing temperature of water (°C), and:

      where $\phi$ is latitude.

      The parameterization for K in 1D models is widely discussed by researchers because the $ K $ value can affect the calculation of the vertical thermal distribution (Sweers, 1970). Previous studies in recent years have proven that the WRF-lake model underestimates the vertical eddy diffusivity, subsequently resulting in energy stored in the upper layers. Enlarging the K value by multiplying with a uniform factor (>1) in the whole vertical profile is usually used to improve the vertical diffusivity and get better performance in water temperature simulations (Gu et al., 2015; Xiao et al., 2016; Xu et al., 2016; Fang et al., 2017). Wang et al. (2019) put forward a better revised scheme in the same deep riverine reservoir as in this study by adopting an enhanced diffusion term (${D}_{{\rm{ed}}}$), referencing to the suggestion of Ellis et al. (1991), Subin et al. (2012), and Fang et al. (2017):

      where $ g $ is the acceleration of gravity, to be 9.8 m s–2, $ \rho $ is density of water (kg m–3) in the corresponding layer, and $ z $ is lake depth under the water surface (m).

      Then the modified K is calculated as follows:

      where $ \alpha $ is an enlarging factor suggested to be 100 by Fang et al. (2017) and Wang et al. (2019) for lakes deeper than 100 m.

    • The inflow-outflow, as an important characteristic of large reservoirs, can affect the vertical mixing and determine the vertical thermal distribution. The larger the flow rate is, the stronger the mixing of the water body and the more unstable the vertical thermal stratification are (Imberger, 1985; Straškraba et al., 1993; Owens, 1998; Nowlin et al., 2004; Li et al., 2016). However, the WRF-lake model does not consider the influence of flow rate on the vertical thermal diffusion coefficient $ K $ in Eq. (6). It is maybe for this reason that although the latest modified version, to our best knowledge, of WRF-lake by Wang et al. (2019) performs better than the original WRF-lake model, it still has an obvious bias during the large flow period. Therefore, a flow rate enlarging term needs to be added into the vertical thermal diffusivity. In this paper, we put forward a new enhanced diffusion term (${K}_{{\rm{v}}}$) to add the influence of flow rate to the $ K $ value calculation. Based on three recognized factors which can affect the vertical thermal stratification, i.e., the flow rate (Straškraba et al., 1993; Owens, 1998; Nowlin et al., 2004; Li et al., 2016), the geometrical proportion of the reservoir (Gorham and Boyce, 1989), and the Brunt–Vasala frequency (Ellis et al., 1991; Subin et al., 2012; Fang et al., 2017; Wang et al., 2019), the function ${K}_{{\rm{v}}}$ = f ($ Q $) needs to satisfy the following conditions:

      1. ${K}_{{\rm{v}}}$ is positively correlated with $ Q $ for larger flow rate results in stronger vertical mixing and thermal diffusivity conditions;

      2. When $ Q $ is zero, ${K}_{{\rm{v}}}$ will be zero, meaning that the reservoir degenerates to a deep lake without throughflow;

      3. ${K}_{{\rm{v}}}$ is also correlated with the topography of the reservoir because, with the same flow rate, a larger and deeper reservoir may have more stable vertical stratification, and a smaller and shallower reservoir may have stronger vertical mixing (Gorham and Boyce, 1989);

      4. The dimension of ${K}_{{\rm{v}}}$ is m2 s–1.

      Based on the four conditions above, we introduced an equation to calculate the ${K}_{{\rm{v}}}$ as follows:

      where $ \beta $ is a dimensionless constant greater than zero, to be 5×10–4 by calibration, $ Q $ is outflow rate (m3 s–1), representing the flow rate in the reservoir, and A is the geometrical proportion of the reservoir (m–0.5), which can describe the topography characteristic and affect the vertical thermal stratification stability, and is calculated as follows (Gorham and Boyce, 1989; Li et al., 2016):

      where S is surface area (m2) and H is the maximum depth of the reservoir (m) (Gorham and Boyce, 1989). Then the modified K introduced in this paper is calculated by:

      The effect of this modification on the vertical diffusion coefficient has been tested in this paper in the following sections 4.1 and 4.2.

    • The Delft3D-Flow model is a 3D model and is a submodule of the Delft3D system, which is developed by Deltares, Holland (Van der Kaaij et al., 2004). The Delft3D-Flow model has been widely used in recent years to study rivers, lakes, and reservoirs and to simulate flows, water quality, waves, water temperature, etc. The Delft3D-Flow model has been proven to be applicable in simulating the thermal processes in large reservoirs (Carrivick et al., 2012; Jiang et al., 2018). For example, Jiang et al. (2018) applied the Delft3D-Flow model to the Jinghong Reservoir, another large riverine reservoir on the Lancang-Mekong River, and got a satisfactory simulation of the water temperature distribution.

      The Delft3D-Flow numerical model solves the 3D shallow water equations, and the equation system consists of the continuity equation, momentum equations, and substance and heat transport equation. The surface heat flux of Delft3D-Flow contains longwave radiation, shortwave radiation, latent heat, and sensible heat, the same as in the WRF-lake model. More details of the mechanisms represented in Delft3D-Flow are in the Appendix.

    3.   Study area, field data and model settings
    • The Lancang-Mekong River is located in the southwest region of China, with a total length of 4880.3 km, and it is called the Lancang River in China and the Mekong River after it enters Laos. The river runs through six countries, namely China, Myanmar, Laos, Thailand, Cambodia, and Vietnam, affecting the livelihood of more than 60 million people (Campbell, 2007; Jiang et al., 2018; Ma et al., 2018; Fan et al., 2015). In recent years, several large reservoirs have been constructed along the Lancang River, e.g., the Nuozhadu Reservoir, the Xiaowan Reservoir, and the Dachaoshan Reservoir. Figure 2 shows the locations of these cascade reservoirs on the middle-lower reach; the Nuozhadu Reservoir is the largest among them and is a typical large, deep riverine reservoir, impounded in 2011 and completed in 2014 (Wang et al., 2019). The length of the reservoir is about 180 km, and the dam is located at 22°38'N, 100°26'E. The Nuozhadu Reservoir is a multi-year regulating reservoir with a total storage capacity of 23.7 billion cubic meters, and the normal storage water level and dead water level of the reservoir are 812 meters and 765 meters above the sea level, respectively.

      Figure 2.  The Lancang-Mekong River basin and the locations of cascade reservoirs on the middle-lower reach.

    • The study period in this paper covers 1 January to 31 December 2015. The Meteorological data were obtained from the China Meteorological Data Sharing Service System (http://cdc.cma.gov.cn/) and drive both the Delft3D-Flow and WRF-lake simulations; the driving data includes shortwave solar radiation (W m–2), longwave radiation (W m–2), air temperature (°C), relative humidity of air (%), wind speed (m s–1), and wind direction (deg). We obtained relative humidity (%) data from the Meteorological Data Sharing Service System and converted it to specific humidity (kg kg–1) to drive the WRF-lake model simulation. These meteorological data were acquired as daily average values. The radiation data were from the Jinghong Weather Station (100°48'N, 22.00°E), and the other meteorological data were from the Lancang Weather Station (99°56'N, 22°34'E). The Delft3D-Flow model additionally required hydraulic data and topographic data. The inflow rates and inflow temperatures, provided by Huaneng Lancang River Hydropower Inc., were set equal to the outflow rates and outflow temperatures of the upstream reservoir, the Dachaoshan Reservoir. The input data of the two models are shown in Fig. 3. The measured temperatures near the Nuozhadu dam, used for the model calibration, were observed by temperature sensors fixed to the dam at different elevations, and these data were also provided by Huaneng Lancang River Hydropower Inc.

      Figure 3.  Monthly averaged inflow-outflow rates and inflow temperature of the Nuozhadu Reservoir in 2015; Daily averaged meteorological data of the Nuozhadu Reservoir in 2015.

    • Model configuration and initialization of the WRF-lake model are based on observed data and calibrations from previous studies:

      (1) The initial depth of the reservoir was set to 194 m, according to the data of Huaneng Lancang River Hydropower Inc.

      (2) The layer division scheme is as follows: According to previous studies, for a lake with depth more than 50 m, a 25-layer discretization is reasonable. The top layer is 0.1 m, and the other 24 layers account for the remaining depths (Gu et al., 2015; Xiao et al., 2016; Wang et al., 2019).

      (3) The lake surface absorption fraction was chosen as 0.4, which is the default in the WRF-lake model and is the same as in Wang et al. (2019).

      (4) The light attenuation coefficient for 1D lake models has been proven in previous studies to be reasonable between 0.13 m–1 and 3 m–1 (Gu et al., 2015; Wang et al., 2019). In our study, 1 m–1 was chosen according to the calibration of Wang et al. (2019) for the Nuozhadu Reservoir.

      (5) The surface roughness length setting is set to be the same as the latest modified scheme used by Subin et al. (2012), which has been proven to be reasonable in recent studies by Huang et al. (2019) and Wang et al. (2019).

      (6) The initial temperature profile was chosen according to the observed vertical temperature distribution provided by Huaneng Lancang River Hydropower Inc.

      (7) The simulation start time and stop time are 1 January and 31 December 2015.

    • Figure 4 is the Delft3D-Flow morphologic grid of the Nuozhadu Reservoir. The horizontal grid contains 330 × 6 cells, and the reservoir is divided into 100 layers vertically. For the 330 × 6 cells, the length is 180 km, and the width is about 500 m at the section Inflow and about 1500 m at the section Dam. The grid spacing is variable according to the horizontal terrain. The grid is finer where the horizontal terrain varies more rapidly. The calculation domain contains two boundaries, including the inflow boundary and the outflow boundary, i.e., the Nuozhadu dam. The initial temperature distribution was also chosen according to the observed vertical temperature distribution. The simulation start time and stop time are 1 January and 31 December 2015, the same as in WRF-lake. The parameter settings were carefully chosen according to previous studies and the calibration in the Nuozhadu Reservoir. The five main parameters are the Chezy coefficient, representing the water bottom roughness, as well as the background eddy viscosity and diffusivity in the horizontal and vertical directions (Chanudet et al., 2012). After the calibration in the Nuozhadu Reservoir, we set the Chezy coefficient to 65 m0.5 s–1, the horizontal and vertical eddy viscosity to be 1 m2 s–1 and 5×10–5 m2 s–1, and the horizontal and vertical eddy diffusivity to be 1 m2 s–1 and 5×10–5 m2 s–1. The value ranges used in this study for the Chezy coefficient, eddy viscosity, and eddy diffusivity, being reasonable to describe the hydraulic process, are consistent with the value ranges used in previous studies (Chow, 1959; Elzawahry, 1985; Tsanis and Wu, 2000; Jiang et al., 2018). The outlet elevation of the Nuozhadu Reservoir, which was controlled by the stop log gate in 2015, was 748 m.

      Figure 4.  The morphologic grid of the Nuozhadu Reservoir.

    • Three model versions, including W0, W1, and D0, were used to simulate the water temperature distribution of the Nuozhadu Reservoir. W0 is the WRF-lake model modified by Wang et al. (2019) and to our best knowledge the latest version of WRF-lake. W1 is the WRF-lake model version modified in this study and is used to test the effect of the enhanced diffusion term $ {K}_{v} $, which was introduced to consider the impact of large flow rate in Eq. (14) (more details in section 2.2).

      D0 represents the Delft3D-Flow model, which is more physically based and can consider the impact of actual topography, inflow-outflow, and longitudinal variation. We used the Delft3D-Flow model to evaluate the reasonableness of the simplifications in WRF-lake when applied to a large, deep riverine reservoir.

    4.   Results and discussions
    • In this section, the effect of parametric scheme optimization of vertical thermal diffusivity ($ K $) in WRF-lake on water temperature simulation is discussed. Some earlier studies modified the parameterization for $ K $ in WRF-lake and proved that an increase in $ K $ could improve the model simulation for deep lakes (Subin et al., 2012; Gu et al., 2015; Xu et al., 2016; Fang et al., 2017). Wang et al. (2019) added a diffusion enhancing term into the model for applications in reservoirs and lakes deeper than 100 m, referring to the suggestions of Ellis et al. (1991), Subin et al. (2012), and Fang et al. (2017). We compared the simulated temperatures by different versions of WRF-lake, including the latest version (W0) (Wang et al., 2019) and the modified version described in section 2.2 of this paper (W1).

      Curves of 2015 surface water temperature simulated by different models, including the two versions of WRF-lake and Delftf3D-Flow, are shown in Fig. 4. Root-mean-square errors (RMSEs) and mean absolute errors (MAEs) are shown in Table 1. The results from Delft3D-Flow are the simulated water temperatures near the Nuozhadu dam. It can be seen that the simulated results of the latest WRF-lake version (Wang et al., 2019) have obvious errors during the large flow rate months, i.e., from May to August. The modification performed in this study eliminated this defect, and the performance was improved significantly because the larger flow rate caused stronger vertical mixing of the water body. Additionally, the downward energy transfer was strengthened, which resulted in lower surface water temperature. The W1 model has a larger $ K $ value than the W0 model in the large flow rate period because of considering the influence of flow rate on vertical diffusivity. Therefore, surface temperatures were simulated lower and closer to the measured data and results of Delft3D-Flow (Fig. 5).

      WRF-lake [by Wang et al. (2019)]WRF-lake (modified)Delft3D-Flow
      Surface water temperatureRMSE (°C)1.501.081.19
      MAE (°C)0.350.150.10
      Vertical water temperature profileRMSE (°C)1.131.011.10
      MAE (°C)0.940.700.60

      Table 1.  The RMSE and MAE of the WRF-lake model version by Wang et al. (2019), the modified WRF-lake model used in this paper, and the Delft3D-Flow model used in surface water temperature and vertical temperature profile simulation near the dam.

      Figure 5.  Simulated surface water temperature in 2015, including the results of the WRF-lake model version by Wang et al. (2019) (blue solid line), the modified WRF-lake model used in this paper (blue dotted line), and the Delft3D-Flow simulation (black solid line), with the observed data (red dots) as reference.

      The vertical profiles of water temperature and diffusion coefficient ($ K $) in the 12 months are shown in Figs. 6(a) and 6(b). The W0 model underestimated vertical mixing in the whole vertical profile during the period from May to August, which was enhanced by large water flow through the reservoir. The revision (W1) of vertical thermal diffusivity in this paper solved this problem, leading the vertical temperature profiles to become more reasonable.

      Figure 6.  (a)Vertical water temperature profiles of the WRF-lake model version by Wang et al. (2019) (blue lines) and the modified WRF-lake model used in this paper (black lines). The dots represent observed data. (b) Diffusion coefficient profiles of the WRF-lake model version by Wang et al. (2019) (blue lines) and the modified WRF-lake model used in this paper (black lines).

    • It has been acknowledged that a vertically 1D model can be applied to a lake-shaped reservoir like the Miyun reservoir (Guo et al., 2022). But it is not clear whether the vertically 1D WRF-lake model can also be applied to a riverine reservoir as long as the Nuozhadu reservoir. Therefore, a 3D model, Delft3D-Flow, was utilized in this study, and the 3D results served as a reference for WRF-lake. Since WRF-lake only considers vertically 1D characteristics, the water temperature variation along the longitudinal direction is discussed in this section.

      Figure 7 shows water temperature isotherms in different months in 2015, including April, July, September, and December, representing spring, summer, autumn, and winter, respectively, simulated by the Delft3D-Flow model. It can be seen that the Nuozhadu Reservoir, with a hydraulic residence time of 180 days, has a significant vertical thermal stratification. This is consistent with previous studies that have found that reservoirs with hydraulic residence times of more than 100 days are usually stably stratified (Straškraba et al., 1993; Owens, 1998; Nowlin et al., 2004; Li et al., 2016). The thermal stratification exists in all four seasons. The water at the surface is always warmer than at the bottom, and the water temperature under 100-m depth remains about 15°C year-round. The longitudinal temperature variation mainly exists in the upstream, within a length of about 50 km, which is 25% of the total length of the reservoir. From the midstream to the dam, the isotherms are relatively horizontal, and the longitudinal water temperature variation is less than 1.5°C with length of 100 km. The isotherms are nearly horizontal near the dam. As a whole, for the large riverine reservoir with hydraulic residence time of 180 days, in more than half of its length, the isotherms are relatively horizontal, and the longitudinal water temperature variation is quite small in all four seasons, which is meaningful for the WRF-lake 1D model application.

      Figure 7.  Simulated water temperature distributions in April, July, September, and December of 2015 by Delft3D-Flow, representing spring, summer, autumn, and winter, respectively.

      The surface heat flux calculation by different versions of WRF-lake was evaluated with the simulated results of Delft3D-Flow as reference. Figures 8a, b, and c show the longwave radiation, latent heat, and sensible heat values simulated by the latest WRF-lake version (Wang et al., 2019) (W0), the WRF-lake modified in this paper (W1), and the Delft3D-Flow model (D0), respectively. The positive values and negative values represent the reservoir releasing energy to and absorbing energy from the atmosphere. The heat flux of Delft3D-Flow is the average in unit area of the whole reservoir, calculated by the ratio of the total surface heat exchange to the total surface area. Table 2 shows the annual average values of these three kinds of heat fluxes. Compared with the Delft3D-Flow calculation, the three kinds of heat fluxes are overestimated by W0, mainly because W0 underestimates the downward transfer of energy in the reservoir, resulting in more heat remaining at the surface and returning to the atmosphere. The W1 model, which considers the effect of flow rate on vertical thermal diffusivity, produced surface heat flux values that were significantly closer to the results of Delft3D-Flow, especially in the large flow rate period, which means that our modification has improved the performance of WRF-lake in surface water–atmosphere interactions. The improvement of surface heat flux calculation appears mainly in summer, i.e., the large flow rate period.

      Figure 8.  Different kinds of heat flux per unit area through the water surface, including net longwave radiation, latent heat, and sensible heat calculated by the WRF-lake model version by Wang et al. (2019) (red lines), the modified WRF-lake model used in this paper (blue lines), and the Delft3D-Flow simulation (black lines). Positive values and negative values represent the reservoir releasing energy to the atmosphere and absorbing energy from the atmosphere

      Net Longwave Radiation (W m–2)Sensible Heat (W m–2)Latent Heat (W m–2)
      Whole yearLarge flow rate periodWhole yearLarge flow rate periodWhole yearLarge flow rate period
      WRF-lake [by Wang et al. (2019)]112.3107.717.414.382.586.2
      WRF-lake (modified)100.382.211.55.660.946.6
      Delft3D-Flow88.484.09.05.950.348.2

      Table 2.  Annual average values and average values of the large flow rate period, i.e., from May to August, for different kinds of heat fluxes per unit area through the water surface, including net longwave radiation, latent heat, and sensible heat calculated by the WRF-lake model version by Wang et al. (2019), the modified WRF-lake model used in this paper, and the Delft3D-Flow simulation. Positive values represent the reservoir releasing energy to the atmosphere.

      Our results show that although W1 significantly outperformed W0, there were still differences between the heat flux simulated by W1 and that of D0, and this was mainly due to the following two reasons: (1) As shown in Fig. 7, there exists obvious longitudinal water temperature variation along the channel in the upstream because the river runs from high latitude to low latitude and the inflow is always colder. The heat flux in the upstream is smaller than that near the dam because of lower surface temperature, lowering the whole reservoir surface heat flux per unit area, which cannot be considered in the WRF-lake model. (2) The temperature difference between inflow and outflow, which is also not considered by WRF-lake but is considered by Delft3D-Flow, can affect the surface heat flux. The inflow is colder than outflow in winter and warmer than outflow in summer, so the heat release simulated by D0 is less in winter and more in summer compared to W1. Therefore, in future studies, for applications in other reservoirs with larger flow rates, the effect of inflow and outflow temperature changes in the WRF-lake model can be investigated further.

    5.   Summary and conclusions
    • In this study, the one-dimensional (1D) lake submodule of the WRF climate system, WRF-lake, was improved by parametric scheme optimization for applications in large, deep riverine reservoirs. The accuracy of water–atmosphere heat flux simulations by the latest version and the revised version of WRF-lake was evaluated for a large, deep riverine reservoir, namely the Nuozhadu Reservoir in the southwest region of China, by comparing to the simulation of the three-dimensional (3D) Delft3D-Flow model, which is a more physically based model and has been previously applied to water temperature distribution simulation in many reservoirs, rivers, and lakes. From the results, the following conclusions were drawn:

      1. The latest version of WRF-lake can describe the surface water temperature of the deep riverine reservoir to some extent, but performs poorly in the large flow rate period. We improved the WRF-lake model by modifying the vertical thermal diffusivity parameterization, and the water temperature simulation in the large flow rate period was significantly improved.

      2. The latest version of WRF-lake overestimates the energy release from the water surface to the atmosphere, and the modification of vertical thermal diffusivity can significantly improve the accuracy in surface water–atmosphere heat flux simulation.

      3. The longitudinal temperature variation and the temperature difference between inflow and outflow, which cannot be considered in the latest version of WRF-lake, can also affect the surface water heat flux.

      Acknowledgements. The authors gratefully acknowledge the financial support from the National Key R&D Program of China (Grant No. 2018YFE0196000) and the National Natural Science Foundation of China (Grant No. 52179069). The valuable discussion with and advice from Professor Danxun Li from the Tsinghua University, Beijing, China are gratefully acknowledged. And we are grateful to the China Meteorological Data Sharing Service System (http://cdc.cma.gov.cn/) for providing the meteorological data, and Huaneng Lancang River Hydropower Inc. for providing the topography, flow rates, and field data, which contributed a lot to our research results reported in this paper.

      Data Availability Statement. All the simulated results, measured data, and code presented in this paper are available upon request to the corresponding author (zhudejun@tsinghua.edu.cn).

    APPENDIX
    • Surface roughness lengths consist of momentum roughness length ($ {z}_{0m} $), heat roughness length (${z}_{0h}$), and water vapor roughness length ($ {z}_{0q} $), which are used to calculate the sensible heat (rah) and latent heat (raw). In the original WRF-lake model, the default values of z0m, z0h, and z0q are set to be identical as 0.0024 m for frozen lakes with resolved snow, and they are set as 0.004 m for frozen lakes. Some previous studies have pointed out that the default surface roughness lengths are too extensive. Subin et al. (2012) provided settings of surface roughness lengths in the ColM-lake 1D lake model as follows: z0m is equal to 0.0024 m for frozen lakes with resolved snow, and 0.001 m for frozen lakes, and:

      where R0 is atmosphere roughness Reynolds number.

      For unfrozen lakes:

      where $ a $ is a dimensionless constant (0.1), $ C $ is the Charnock coefficient, $ {u}_{*} $ is friction velocity, $ v $ is kinematic viscosity of air, and $ g $ is the acceleration of gravity.

      The settings of surface roughness lengths as illustrated above have been very extensively applied in the WRF-lake model (Xu et al., 2016; Huang et al., 2019; Wang et al., 2019) by researchers after being proposed by Subin et al. (2012). In this paper, we also followed this parametric scheme.

    • The continuity equation is expressed in the $ \xi $-direction and $ \eta $-direction as follows (Deltares, 2003):

      where $ \zeta $ is the water level (m), $ \sqrt{{G}_{\xi \xi }}\sqrt{{G}_{\eta \eta }} $ are used to transfer the curvilinear coordinates to the rectangular coordinates (m), t is time (s), d is depth (m), U and V are respectively depth-averaged speed in the $ \xi $-direction and $ \eta $-direction (m s–1), and $ Q $ is the source term including water release, evaporation, and precipitation of each unit.

      The momentum equations in the $ \xi $-direction and $ \eta $-direction are as follows (Deltares, 2003):

      where $ u $, $ v $, and $ \omega $ are respectively speed in the $ \xi $-, $ \eta $-, and $ \sigma $-directions (m s–1), $ {\rho }_{0} $ is the referential density of water (kg m–3), $ f $ is the Coriolis parameter (s–1), $ {P}_{\xi } $ and $ {P}_{\eta } $ are the gradient-based hydrostatic pressures (kg m–2 s–2), $ {F}_{\xi } $and $ {F}_{\eta } $ are the imbalances existing in horizontal Reynold stresses (m s–2), $ {v}_{V} $ is the vertical eddy viscosity (m2 s–1), and $ {M}_{\xi } $ and $ {M}_{\eta } $ are the source and sink terms of momentum.

      The heat transport process is described by the advection diffusion equation in three directions as follows (Delft 2003):

      where $ I $ is the concentration of heat (J m–3), $ {D}_{H} $ and $ {D}_{V} $ are the horizontal diffusion coefficient (m s–1) and the vertical diffusion coefficient (m2 s–1), $ {\lambda }_{d} $ (s–1) is the first order decay course, and $ S $ is the source and sink term per unit area caused by withdrawal and release of water and water surface heat exchange (J m–2 s–1).

    • Figure 6.  (Continued)

Reference

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return