Environment and Interdisciplinary Development

Performance Analysis of LARS-WG and SDSM Models in Downscaling the Output of CMIP6 Models in The Urmia Lake Watershed

Document Type : Original Article

Authors

Navid Dehghani 1
Massoud Goodarzi 2
Mahshid Karimi 3

1 Department of Watershed Management, Soil Conservation and Watershed Management Research Institute (SCWMRI), AREEO, Tehran, Iran

2 Department of Drought and Climate Change, Soil Conservation and Watershed Management Research Institute (SCWMRI), AREEO, Tehran, Iran

3 Department of Soil Conservation and Watershed Management, Kermanshah Research center for Agiculture and Natural Resources, AREEO, Kermanshah, Iran

https://doi.org/10.22034/envj.2025.535309.1523
Abstract
Introduction: The output of general circulation models (GCMs) lacks the spatial and temporal accuracy required for regional and local studies due to the low spatial resolution of the grid. This has led to the development of regional models and statistical and dynamical downscaling. Among the statistical methods, LARS-WG and SDSM models `are considered to be the most convenient and reliable downscaling tools.
Methods and Materials: In this study, the performance of these two models were evaluated using downscaling the output of CMIP6 models and simulating temperature and precipitation variables in the Urmia Lake basin. The meteorological stations studied included 7 synoptic stations with a statistical period of 30 years corresponding to the base period of the models (1985-2014). The error coefficients including MSE, RMSE, MAE and R2 were also used to evaluate the performance of the models.
Results: Both LARS-WG and SDSM models have high ability in simulating temperature and precipitation variables in the studied area. However, their accuracy was not the same at different stations and also for different climatic variables. Based on the results, both models have lower accuracy in simulating precipitation than temperature, which could be due to the complexity of the precipitation process and its nature. The SDSM model has the lowest error in simulating temperature and precipitation data in the study basin. Although the LARS-WG model also has a good ability to simulate temperature and precipitation data for the small scale and has performed better in some stations, but its capability is not as good as the SDSM model. In SDSM model, the downscaling operation is performed by creating a regression relationship between predictors and predictors at a station, but in the LARS-WG model, independent and large-scale atmospheric variables do not play a direct role in simulating the data. Rather, the model initially analyzes the observational data to determine their parameters and statistical properties. Furthermore, in line with the type of future changes in large-scale climate variables, it changes the statistical parameters of the observational data and attempts to reproduce the data in future periods. Considering the error metrics and comparing the two models under study with each other, one cannot definitely prefer one over the other. 
Discussions: Due to the type of simulation process and the combined structure of the SDSM model in data downscaling and the direct use of general atmospheric circulation models and large-scale NCEP and ERA5 data in the sixth report, this model has greater accuracy in simulating data in the studied basin. On the other hand, the LARS-WG model is superior due to the simplicity of the model structure, the input data to the model, the need for less skill, and the speed of operation, and it gives the user more flexibility. However, the SDSM model has a more complex process and requires more accuracy and time, as well as relatively high user expertise. In addition to climatic variables, this model can also be used for hydrological and environmental variables, while the LARS-WG model is only applicable to temperature, precipitation, radiation, and evaporation variables. But on the other hand, for the LARS-WG model, considering the climatic characteristics of the studied region, new climate change scenarios can be defined, which can be useful in the application of these models in climate change areas. Overall, it can be concluded that these models, despite their differences, can produce the statistical behavior of climate data from a weather station in terms of mean, standard deviation, etc., which are identical to the statistical behavior of observational data, and none of the models has absolute superiority over the other. Given that in any region, before implementing climate change models, it is necessary to downscale and evaluate the performance of the models, the results of this research can be used to verify the output of CMIP6 models in predicting climate variables in future periods in different climates.

Keywords

Climate Change
Urmia Lake
Downscaling LARS-WG
SDSM

Subjects

1.    Aghashahi, M., Ardestani, M., Niksokhan, M.H. and Tahmasbi, B., 2012. introducing and comparisons of Lars-wg and SDSM modelsfor downscaling environmental parameters and climate change studies, 6th national conference and exhibitions of environmental engineering.
2.    Babaousmail, H., Hou, R., Ayugi, B., Ojara, M., Ngoma, H., Karim, R., Rajasekar, A. and Ongoma, V., 2021. Evaluation of the Performance of CMIP6 Models in Reproducing Rainfall Patterns over North Africa. Atmosphere, 12, 475.
3.    Carter, T.R., Parry, M.L., Harasawa, H. and Nishioka, S., 1994. IPCC technical guidelines for assessing climate change impacts and adaptions, IPCC Special Report to Working Group II of IPCC, London.
4.    Dibike, Y.B. and Coulibaly, P., 2005. Hydrologic impact of climate change in the Saguenay Watershed: Comparison of Ownscaling Methods and Hydrologic Models. Journal of Hydrologic, 307: 145–163.
5.    Goodarzi, M., Hosseini, A. and Mesgari, E., 2016. Climate Models, Azarkelk press, Zanjan.
6.   Goodarzi, M., Salahi, B. and Hosseini, S.A., 2016. Performance Analysis of LARS-WG and SDSM Downscaling Models in Simulation of Climate Changes in Urmia Lake Basin. jwmseir; 9 (31):11-22. UR:http://jwmsei.ir/article-1-457-en.html
7.    Hafezi Moghaddas, N., Lashkaripour, G. and Parsaei, R., 2024. Performance analysis of CMIP6 models in projection of temperature and precipitation changes in the Chahnimeh area of Sistan and Baluchistan province. Journal of Climate Research, 1402(56), 165-178.
8.    Hashmi. M.Z., Shamseldin. A.Y. and Melville, B.W., 2010. Comparison of SDSM and LARS-WG for simulation and downscaling of extreme precipitation events in a watershed, Stoch Environ Res Risk Assess.
9.    Hosseini, A., 2015, a study on climate change impacts on the changes of surface runoffin Urmia Lake basin, a PhD dissertation, Physical geography, Mohaqegh Ardabili University, 196pp.
10. Hu, T.S., Lam, K.C. and Ng, S.T., 2001. River flow time series prediction with a range dependent neural network. Hydrological Science Journal, 46: 729-745.
11.Karamouz, M., Ramezani, F. and Razavi, S., 2006. application of Artificial Neural network, in predicting long-term precipitation using Meteorologic signals, 7th international congress on civil engineering, Tehran. 11 pp.
12.Khan, M.S., Coulibaly, P. and Dibike. Y., 2006. Uncertainty analysis of statistical downscaling method. Journal of Hydrology, 319: 357-382.
13.Kilsby, C.G., Jones, P.D., Burton, A., Ford, A.C., Fowler, H.J., Harpham, C., James, P., Smith, A. and Wilby, R.L., 2007. A daily weather generator for use in climate change studies. Environmental Modelling and Software, 22: 1705–1719.
14.Lin, J.Y., Cheng,C.T. and Chau, K.W., 2006. Using support vector machines for long-term discharge prediction. Hydrological Science Journal, 51: 599-612.
15.Majdi, F.,  Hosseini, S.A., Karbalaee, A., Kaseri, M. and Marjanian, S., 2022. Future projection of precipitation and temperature changes in the Middle East and North Africa (MENA) region based on CMIP6. Theor Appl Climatol 147, 1249–1262.
16.Mesgari, M., Hosseini, S.A., Hemmesy, M.S., Houshyar, M. and Golzari Partoo, L., 2022. Assessment of CMIP6 models’ performances and projection of precipitation based on SSP scenarios over the MENAP region. Journal of Water and Climate Change, 13 (10): 3607–3619.
17. Oji, R., 2018. Comparison of Multi-site and Single-site Daily Precipitation and Temperature Extremes Downscaling (Case Study: Southern Coast of the Caspian Sea). Journal of the Earth and Space Physics, 44(2), 397-410. doi:10.22059/jesphys.2017.234927.1006908
18.Peng, S., Wang, C., Li, Z., Mihara, K., Kuramochi, K., Toma, Y. and Hatano, R., 2023.Climate change multi-model projections in CMIP6 scenarios in Central Hokkaido, Japan. Sci Rep, 13(1):230.
19.Racsko, P., Szeidl, L. and Semenov, M., 1991. A serial approach to local stochastic weather models. Ecological Modeling, 57, 27-41.
20.Samadi, Z. and Massah Bavani, A., 2008. Introducing artifitial neural network and SDSM methods for downscaling temperature and precipitation data, 3rd conference on Iranian water resources, university of Tabriz, 9p.
21.Sedaghat Kerdar, A. and Fatahi, E., 2008. Drought Early Warning Methods over Iran. Geography and Development, 6(11), 59-76. doi: 10.22111/gdij.2008.1616Salahi, B., Goudarzi, M. and Hosseini, S. A. (2016). Predicting the temperature and precipitation changes during he 2050s in Urmia Lake Basin. Watershed Engineering and Management, 8(4), 425-438. doi: 10.22092/ijwmse.2016.107179.
22.Semenov, M., Brooks, R., Barrow, E. and Richardson, C., 1998. Comparison of the WGEN and LARS-WG stochastic weather generators for diverse climates. Clim. Res., 10:95-107.
23.Semonov, M.A. and Stratonovitch, P., 2010. Use of multi-model ensembles from global climate models for assessment of climate change impacts, Climate Research, 41: 1-14.
24.Shamsipour, A.A., 2013. climate modeling, concepts and methods, Tehran university press, 294 pp.
25.Sharma D., Gupta A.D., Babel M.S.  2007. Spatial disaggregation of bias-corrected GCM precipitation for improved hydrologic simulation: Ping River Basin, Thailand, Hydrol. Earth Syst. Sci. 11: 1373–1390.
26. Sobhani, B., Eslahi, M. and Babaeian, I., 2017. Comparison of statistical downscaling in climate change models to simulate climate elements in Northwest Iran. Physical Geography Research, 49(2), 301-325. doi: 10.22059/jphgr.2017.62847
27. Sunyer, M.A., Hundecha, Y., Lawrence, D., Madsen, H., Willems, P., Martinkova, M., Vormoor, K., Bürger, G., Hanel, M., Kriaučiūnienė, J., Loukas, A., Osuch, M. and Yücel, I., 2015. Inter-comparison of statistical downscaling methods for projection of extreme precipitation in Europe, Hydrol. Earth Syst. Sci., 19:1827-1847.
28. Wigley, T.W.L., Jones, P.D., Briffa, K.R. and Smith, G., 1990. Obtaining sub-grid scale information from coarse resolution general circulation model output, J. Geophys. Res. 951: 1943–1953.
29. Wilby, R.L., Dawson, C.W. and Barrow, E.M., 2002. SDSM- a decision support tool for the assessment of regional climate change impacts, Environmental Modeling & Software, 17: 147-159.
30. Wilby, R.L. and Harris. I., 2006. A frame work for assessing uncertainties in climate change impacts: low flow scenarios for the River Thames, UK. Water Resour. Res. 42:10 p.
31. Wilby, R.L., Tomlinson, O.J. and Dawson, C.W., 2007. Multi-site simulation of precipitation by condition resampling. Journal of climate Research, 23: 183-194.
32. Wilks, D.S., 1992. Adapting stochastic weather generation algorithms for climate change studies. Climate Change.. 22: 67-84.
33. Wilks, D.S. and Wilby, R.L., 1999. The weather generation game: a review of stochastic weather models. Progress in Physical Geography. 23: 329-357.
34. Zareian, M.J., Dehban, H. and Gohari, A., 2023. Evaluation of the Accuracy of CMIP6 Models in Estimating the Temperature and Precipitation of Iran Based on a Network Analysis. Water and Irrigation Management, 12(4), 783-797. doi:  10.22059/jwim.2022.345975.1006
Environment and Interdisciplinary Development
Volume 10, Issue 88
Summer 2025
Pages 98-111

Home

Submit Manuscript

Contact Us