Investigating effects of better soil moisture initialization on forecasting capability with WRF model

thumbnail.default.alt
Tarih
2023-01-19
Yazarlar
Haliloğlu, Şule
Süreli Yayın başlığı
Süreli Yayın ISSN
Cilt Başlığı
Yayınevi
Graduate School
Özet
This thesis aims to analyze the effect of soil moisture on precipitation over Turkey by using of numerical weather prediction model (The Weather Research and Forecasting: WRF). The model domain covers most of the Black Sea, the Eastern Mediterranean Sea, and surrounding countries. In this study, satellite based soil moisture (ASCAT : Advanced Scatterometer) and precipitation (MSWEP : Multi-Source Weighted-Ensemble Precipitation) observations have been analyzed to define WRF simulation periods. WRF simulations have been accomplished in 3-km horizontal resolution forcing with ERA5 Re-analysis data. Analyses of Multi-annual observational datasets of precipitation and soil moisture for the period 1992-2016 pronounced that years of 2007 and 2015 are appropriate for deeper analysis. Anomalies of soil moisture and precipitation analyses exhibit dry and wet years which are 2007 and 2015, repectively. Based on these analyses, WRF simulations have been designed as three months (May-June-July), two months (June-July), and one month (July) for these two years to detect effects of soil moisture. This methodology has been applied to understand the effect of soil moisture on precipitation by different initialization timing. Selected WRF outputs, the maximum and minimum temperature, soil moisture, and precipitation have been compared with ERA5 Re-analysis. This comparison revealed that the three-month WRF simulation has cold bias up to 5oC over the western and central Anatolian region for maximum temperature in July while the minimum temperature bias reaches 6oC over the southeast in the one-month simulation of 2007. In 2015, all three simulations deliver lower minimum temperatures, approaching 4oC in the southeast, than ERA5 re-analysis. It is observed that the precipitation difference between simulations and reanalysis data is up to 80 mm. All monthly simulations have been compared with each other over the selected subregions which are Aegean Region, Central Anatolia Region and Southeast Region. It is detected that the temperature results are improved in one-month 2007 simulation by 3oC due to the initial condition changes in the soil moisture. For the year of 2015, wet initial conditions of the soil produced 2oC colder maximum temperatures over Central Anatolian Region in July. For the selected regions, time series analyses of precipitation and soil moisture are conducted and it exposes that the effect of initial soil moisture on precipitation is larger in the dry year.
Açıklama
Thesis (M.Sc.) -- İstanbul Technical University, Graduate School, 2023
Anahtar kelimeler
Soil moisture, Toprak nemi, Simulation, Benzetim, Observational learning, Gözleyerek öğrenme, Estimation methods, Tahmin yöntemleri
Alıntı