Verification of operational Universal Thermal Climate Index forecasts for Slovenia

Kuzmanović, Danijela (author), Skok, Gregor (mentor), Banko, Jana (comentor)

The Universal Thermal Climate Index (UTCI ) can be described as the air temperature of the reference environment which causes the same human body response as an actual state of environment with a given combination of meteorological parameters (air temperature, wind, relative humidity and radiation). It is one of the thermal indices based on a heat balance model, taking into account the meteorological factors and the influence of the clothing. The UTCI has been widely applied in research and recently it was also implemented in ALADIN model and used in Slovenia. The main goal of this master thesis was a verification of operational UTCI forecasts for Slovenia. The Slovenian Environment Agency (ARSO) has provided us with archived operational outputs of ALADIN’s UTCI forecasts for the period 2013-2018 and for the year 2020. We verified it based on the measured meteorological data, where the UTCI values were calculated with BioKlima model. The measured meteorological data came from 42 stations in Slovenia, where the radiation measurements were preformed. The verification was done for the dataset from all stations together and separately for 9 selected stations. The results show that the model on average overestimates UTCI values. The overestimation is more pronounced in the morning, when the hourly average mean error (ME) reaches 8 °C. The daily average ME is 2.57 °C and daily average MAE is 5.02 °C. There is a small percentage of cases with the very large error (up to 40 °C). The secondary goal of the master thesis was to improve the operational ALADIN forecasts of UTCI. We used two methods, linear regression (LR) and neural network (NN). Both methods have successfully managed to reduce the ME and MAE. Also, according to MAE, NN performed better than LR, but that difference is rather small.

• Language: English
• Keywords: UTCI, ALADIN, BioKlima, verification, thermal indices, heat stress, cold stress
• Work type: Master's thesis/paper
• Typology: 2.09 - Master's Thesis
• Organization: FMF - Faculty of Mathematics and Physics
• Year: 2022
• PID: 20.500.12556/RUL-143822
• COBISS.SI-ID: 137599747
• Publication date in RUL: 13.01.2023

Secondary language

• Language: Slovenian
• Title: Verifikacija operativnih napovedi univerzalnega toplotnega klimatskega indeksa v Sloveniji

Abstract

Univerzalni toplotni klimatski indeks (UTCI) lahko opišemo kot temperaturo zraka referenčnega okolja, ki povzroči enak odziv človeškega telesa kot dejansko stanje okolja z dano kombinacijo meteoroloških parametrov (temperatura zraka, veter, relativna vlažnost in sevanje). Je eden izmed toplotnih indeksov, ki temelji na modelu toplotne bilance. Upošteva meteorološke dejavnike in vpliv oblačil. UTCI se široko uporablja v raziskavah in pred kratkim je bil implementiran tudi v model ALADIN. Glavni cilj magistrske naloge je bila verifikacija operativnih napovedi UTCI za Slovenijo. Verifikacija modela numerične vremenske napovedi je pomemben proces za uspešen razvoj in izboljšanje napovedi z obsežno analizo njegovih prednosti in slabosti. Napovedi modela UTCI smo pridobili od Agencije RS za okolje (ARSO). ARSO nam je posredoval arhivirane operativne rezultate napovedi modela ALADIN za obdobje 2013-2018 in za leto 2020. Verificirali smo jih na osnovi izmerjenih meteoroloških podatkov, kjer so vrednosti UTCI izračunane v modelu BioKlima. Izmerjeni meteorološki podatki so iz 42 postaj v Sloveniji, kjer so bile opravljene meritve sevanja. Verifikacija je narejena za nabor podatkov z vseh postaj skupaj in posebej za devet izbranih postaj. Rezultati kažejo, da model v povprečju precenjuje vrednosti UTCI. Precenjenost je izrazitejša zjutraj, ko urno povprečena srednja napaka (ME) doseže največ 8 °C. Dnevna povprečna napaka (ME) je 2.57 °C, dnevna povprečna srednja absolutna napaka (MAE) je 5.02 °C. Rezultati vsebujejo majhen odstotek primerov z napako, ki bistveno odstopa od povprečja (do 40 °C). Drugotni cilj magistrske naloge je bil izboljšati operativne ALADIN-ove napovedi UTCI. Za izboljšanje napovedi in zmanjšanje napake smo uporabili dve metodi, linearno regresijo (LR) in nevronsko mrežo (NN). Obe metodi sta uspešno uspeli zmanjšati ME in MAE. Glede na MAE se je NN izkazal boljše kot LR, kot smo pričakovali.

• Keywords: UTCI, ALADIN, BioKlima, verifikacija, termalni indeks, toplotni stres, hladni stres