Prediction of Stable Isotopes (18O and 2H) in the Bangkok Metropolitan Area’s Precipitation Using an Artificial Neural Network †
Faculty of Environment and Resource Studies, Mahidol University, Nakhon Pathom 73170, Thailand
*
Author to whom correspondence should be addressed.
†
Presented at the 5th International Electronic Conference on Atmospheric Sciences, 16–31 July 2022; Available online: https://ecas2022.sciforum.net/ .
Environ. Sci. Proc. 2022, 19(1), 32; https://doi.org/10.3390/ecas2022-12792
Published: 14 July 2022
(This article belongs to the Proceedings of The 5th International Electronic Conference on Atmospheric Sciences)
Abstract
:The role of local (wind speed, potential evaporation, vapor pressure, air temperature, and precipitation amount) and regional parameters (teleconnection indices such as Indian Ocean Dipole (IOD), Bivariate ENSO index (BEST), North Atlantic Oscillation (NAO), Southern Oscillation index (SOI), and Quasi Biennial Oscillation (QBO) on the stable isotope content in the precipitation in Bangkok was investigated. First, a simple artificial neural network (ANN) and a Deep Learning Neural Network (DNN) were used to predict the stable isotope content in precipitation. Second, studying the fractional importance of various parameters on the stable isotope content of precipitation demonstrated that among the local and regional parameters, precipitation amount and potential evaporation (local) and the BEST teleconnection index (regional) had dominant roles in controlling the stable isotope content of the precipitation.
1. Introduction
Thailand is a country in southeast Asia located in the tropical climate zone between the Pacific and Indian oceans. According to the Köppen climate classification, Thailand is mainly covered by two dominant climate zones, tropical savanna (Aw) and tropical monsoon (Am) [1]. Thailand’s climate is mainly controlled by the seasonal variations of moisture transfer from the Indian and Pacific oceans, through the northeast monsoon from the Pacific Ocean and the southwest monsoon from the Indian Ocean [2] (Figure 1). The northeast monsoon (NE) affects Thailand from mid-October to mid-February, and the role of the Pacific Ocean is dominant during this period. The southwest monsoon (SW) influences Thailand from mid-May to mid-October, and the precipitation moisture originates from the Indian Ocean [2].
The spatial and temporal variations in precipitation moisture sources normally influence the stable isotope content of precipitation. Studying and monitoring the stable isotopes (18O and 2H) in the precipitation and other elements of the hydrological cycle such as groundwater or surface water resources can provide valuable information for many areas of the hydrological, hydrogeological, and climatological sciences [4]. Stable isotope sampling in precipitation across Thailand commenced in 1968 when the first Global Network of Isotopes in Precipitation (GNIP) station was established in the Bangkok metropolitan area. Bangkok is the capital and most populated metropolitan area of Thailand, located in the Chao Phraya delta in the central part of this country.
The aim of this investigation is to study comprehensively the role and fractional importance of various regional (teleconnection indices) and local parameters on the stable isotope content of the precipitation in the Bangkok metropolitan area. In addition, the stable isotope content in Bangkok precipitation was simulated using artificial neural network techniques. Bangkok GNIP station is one of the major stations in the world with over 47 years of monthly stable isotope data sets, which makes it suitable for the current study.
2. Materials and Methods
In the following survey, the stable isotopes (18O and 2H) in the precipitation samples from 1968 to 2015 at the Bangkok station were studied. The stable isotopes were all presented in delta notation δ relative to the Vienna Standard Mean Ocean Water (VSMOW) standard and presented in ‰ units with the analytical uncertainties of 0.1‰ and 1‰ for the 18O and 2H isotopes, respectively. The local parameters including wind speed and potential evaporation were downloaded from the NOAA website (http://www.cpc.ncep.noaa.gov (accessed on 30 August 2022)) [5]. However, other parameters such as the precipitation amount, vapor pressure, and air temperature were also presented in the same file as the stable isotopes in the GNIP data set for Bangkok. In addition to the local factors, the role of regional parameters (teleconnection indices) on the stable isotope content in Bangkok precipitation were studied. The Southern Oscillation Index (SOI), Bivariate ENSO (BEST), the Quasi-Biennial Oscillation (QBO), the North Atlantic Oscillation (NAO), the Indian Ocean Dipole (IOD), and the El Nino-Southern Oscillation (ENSO) are the main teleconnection indices influencing climate in southern part of Asia including Thailand [6,7]. These teleconnection indices data are available for free from the NOAA website [5,8].
To simulate the stable isotope content of precipitation as well as determine the fractional importance of various parameters in controlling the stable isotopes content of precipitation, artificial neural network techniques were used. The ability and the accuracy of a simple artificial neural network (ANN) as well as a Deep Learning Neural Network (DNN) to forecasts stable isotopes content in Bangkok precipitation were investigated. The ANN model unlike conventional statistical techniques is applicable to problems that contain complicated nonlinear interactions, which makes it a reliable method to forecast stable isotope content in precipitation [9,10,11,12]. To develop predictive models, local (wind speed, potential evaporation, vapor pressure, air temperature, and precipitation amount) as well as regional parameters (teleconnection indices such as IOD, BEST, NAO, SOI, and QBO) were used to predict the stable isotope content in Bangkok precipitation. In addition, to involve the role of moisture sources in the models, the stable isotope data set was classified into three groups, based on the month of precipitation sampling, SW monsoon (group1), NE monsoon (group2), and transition (group3). Finally, the role and fractional importance of each independent local and regional parameter in the final models was estimated by the DNN and ANN models.
3. Results and Discussion
The stable isotope signatures in the precipitation of Bangkok was studied, and the stable isotope content was simulated by a simple ANN model and a more complicated DNN model. To simulate the stable isotopes in the precipitation, the local and regional parameters influencing the stable isotopes content in precipitation were entered into the DNN and ANN models as input data. Comparing the simulated and real isotope data (Figure 2), both DNN and ANN models simulated the stable isotope content with acceptable accuracy. To quantify the accuracy of the simulated stable isotope content, the correlation between the real and simulated stable isotopes was investigated, and the coefficient of determination (R2) was calculated for both models (Figure 3). The results showed that the R2 values for both isotopes (18O and 2H) and for both models were approximately the same ranging from 62% to 66%. These are not very high R2 values, but the accuracy is acceptable for the developed models. To achieve more accurate models, the role of other parameters including cloud top pressure (CTP), cloud top temperature (CTT), and the role of inland/continental moisture sources should also be considered. The CTP and CTT data sets were not available for the entire Bangkok stable isotopes data sets (only for 15 years from 2000 to 2015). However, no comprehensive investigations have been conducted on the role of inland/continental moisture sources on the stable isotope content of the precipitation in Thailand yet. Therefore, these parameters could not be involved in the developed models in this study.
In addition to the simulation of the stable isotopes in precipitation, the fractional importance of each local and regional parameter influencing the stable isotopes in precipitation were also calculated by both the DNN and ANN models (Table 1). According to the outputs of the models, the potential evaporation and precipitation amounts were the main local parameters influencing the stable isotopes in precipitation. The BEST teleconnection index was the main regional parameter influencing the stable isotope content of the precipitation in Bangkok. Other local and regional parameters also influenced the stable isotope content of the precipitation, but they had a weak role compared to the mentioned factors.
4. Conclusions
The results of this study revealed that both the DNN and ANN models provided approximately the same level of accuracy to forecast the stable isotope content in Bangkok’s precipitation. Furthermore, studying the role of various local and regional parameters also showed that the precipitation amount and potential evaporation (among the local parameters) and the BEST index (among the regional parameters) were the dominant factors controlling the stable isotope content of the precipitation in Bangkok.
Author Contributions
Conceptualization, M.H. and N.P.; methodology, N.P.; software, M.H.; validation, M.H.; formal analysis, M.H.; investigation, M.H.; resources, N.P.; writing—original draft preparation, M.H.; supervision, N.P.; project administration, N.P. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by the Faculty of Environment and Resource Studies, Mahidol University, grant number MU-PD-2021-13.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
Data will be available on request.
Acknowledgments
Special thanks to the National Oceanic and Atmospheric Administration (NOAA), U.S. Department of Commerce for providing us with teleconnection indices data sets, as well as the Global Network of Isotopes in Precipitation (GNIP) for providing the precipitation isotope data set for Bangkok. The first author also acknowledges the postdoctoral fellowship (no. MU-PD-2021-13) from the Faculty of Environment and Resource Studies, Mahidol University. The authors also express their gratitude to Supaporn Buajan for her help during this work.
Conflicts of Interest
The authors declare no conflict of interest.
References
- Khedari, J.; Sangprajak, A.; Hirunlabh, J. Thailand climatic zones. Renew. Energy 2002, 25, 267–280. [Google Scholar] [CrossRef]
- Manisan, V. Geography and Climatology in Every Season of Various Parts in Thailand; Hachette Book Group: Bangkok, Thailand, 1995. [Google Scholar]
- Laonamsai, J.; Ichiyanagi, K.; Kamdee, K.; Putthividhya, A.; Tanoue, M. Spatial and temporal distributions of stable isotopes in precipitation over Thailand. Hydrol. Process. 2020, 35, e13995. [Google Scholar] [CrossRef]
- Clark, I.D.; Fritz, P. Environmental Isotopes in Hydrogeology; CRC Press (Lewis Publishers): Boca Raton, FL, USA, 1997; ISBN 1566702496. [Google Scholar]
- NOAA. Available online: http://www.cpc.ncep.noaa.gov (accessed on 30 August 2022).
- Pong, L.; Xuhui, L.; Uma, W. The Role of Teleconnection Indices in Precipitation Amount Variations of South Part of Asia; Chinese Academy of Sciences: Beijing, China, 2002. [Google Scholar]
- Ichiyanagi, K.; Yamanaka, M. Interannual variation of stable isotopes in precipitation at Bangkok in response to El Ñino Southern Oscillation. Hydrol. Process. 2005, 19, 3413–3423. [Google Scholar] [CrossRef]
- NOAA. Available online: https://www.ncdc.noaa.gov (accessed on 30 August 2022).
- Schroeter, B. Artificial Neural Networks in Precipitation Nowcasting: An Australian Case Study; Springer International Publishing: Cham, Switzerland, 2016; Volume 628, pp. 325–339. ISBN 978-3-319-28493-4. [Google Scholar]
- Purnomo, H.D.; Hartomo, K.D.; Prasetyo, S.Y.J. Artificial Neural Network for Monthly Rainfall Rate Prediction. IOP Conf. Ser. Mater. Sci. Eng. 2017, 180, 12057. [Google Scholar] [CrossRef]
- Gholizadeh, M.H.; Darand, M. Forecasting Precipitation with Artificial Neural Networks (Case Study: Tehran). J. Appl. Sci. 2009, 9, 1786–1790. [Google Scholar] [CrossRef]
- Mislan, H.; Hardwinarto, S.; Sumaryono, M.A.; Aipassa, M. Rainfall Monthly Prediction Based on Artificial Neural Network: A Case Study in Tenggarong Station, East Kalimantan-Indonesia. Procedia Comput. Sci. 2015, 59, 142–151. [Google Scholar] [CrossRef]
Figure 1.
The main air masses influencing Thailand, the direction of SW and NE monsoons, and the study area location (adapted from Laonamsai et al. [3]).
Figure 1.
The main air masses influencing Thailand, the direction of SW and NE monsoons, and the study area location (adapted from Laonamsai et al. [3]).
Figure 2.
The comparison between the real and simulated stable isotopes data using the DNN and ANN models on Bangkok’s precipitation.
Figure 2.
The comparison between the real and simulated stable isotopes data using the DNN and ANN models on Bangkok’s precipitation.
Figure 3.
The regression correlation between the real and simulated stable isotopes data of Bangkok’s precipitation and R2 score values.
Figure 3.
The regression correlation between the real and simulated stable isotopes data of Bangkok’s precipitation and R2 score values.
Table 1.
Calculated fractional importance of the main local and regional parameters influencing δ18O and δ2H in Bangkok’s precipitation.
Table 1.
Calculated fractional importance of the main local and regional parameters influencing δ18O and δ2H in Bangkok’s precipitation.
| Parameter | ANN-δ18O | DNN-δ18O | ANN-δ2H | DNN-δ2H |
|---|---|---|---|---|
| IOD | 0.04 | 0.06 | 0.05 | 0.07 |
| NAO | 0.07 | 0.06 | 0.06 | 0.03 |
| QBO | 0.05 | 0.05 | 0.03 | 0.03 |
| SOI | 0.07 | 0.04 | 0.07 | 0.05 |
| BEST | 0.15 | 0.12 | 0.16 | 0.15 |
| Wind speed | 0.05 | 0.06 | 0.05 | 0.04 |
| Evaporation | 0.23 | 0.26 | 0.25 | 0.27 |
| Vapor pressure | 0.07 | 0.08 | 0.08 | 0.09 |
| Air temperature | 0.10 | 0.07 | 0.06 | 0.05 |
| Precipitation amount | 0.17 | 0.15 | 0.16 | 0.19 |
| Season | 0.03 | 0.05 | 0.04 | 0.03 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Heydarizad, M.; Pumijumnong, N. Prediction of Stable Isotopes (18O and 2H) in the Bangkok Metropolitan Area’s Precipitation Using an Artificial Neural Network. Environ. Sci. Proc. 2022, 19, 32. https://doi.org/10.3390/ecas2022-12792
AMA Style
Heydarizad M, Pumijumnong N. Prediction of Stable Isotopes (18O and 2H) in the Bangkok Metropolitan Area’s Precipitation Using an Artificial Neural Network. Environmental Sciences Proceedings. 2022; 19(1):32. https://doi.org/10.3390/ecas2022-12792
Chicago/Turabian StyleHeydarizad, Mojtaba, and Nathsuda Pumijumnong. 2022. "Prediction of Stable Isotopes (18O and 2H) in the Bangkok Metropolitan Area’s Precipitation Using an Artificial Neural Network" Environmental Sciences Proceedings 19, no. 1: 32. https://doi.org/10.3390/ecas2022-12792
APA StyleHeydarizad, M., & Pumijumnong, N. (2022). Prediction of Stable Isotopes (18O and 2H) in the Bangkok Metropolitan Area’s Precipitation Using an Artificial Neural Network. Environmental Sciences Proceedings, 19(1), 32. https://doi.org/10.3390/ecas2022-12792
