Abstract
Teleconnection events can influence normal regional weather patterns and affect weather forecast accuracy. To improve the forecast ability, the relationship between main teleconnections such as El Niño–Southern Oscillation (ENSO), Indian Ocean Dipole (IOD), Madden–Julian Oscillation (MJO), and climate variables (rainfall, maximum and minimum surface temperature, vertical mixing ratio, and vertical maximum temperature) was established using lag correlation coefficient and t-test methods. The results reveal moderately significant correlations between El Niño, positive IOD and rainfall, and vertical mixing ratio, which can be associated with lower-than-usual rainfall. The coincidence between El Niño and positive IOD events can worsen drought. Even though the MJO and regional weather correlations were significant, the magnitude of correlation coefficients was negligible. In addition, the spatiotemporal distribution of ENSO shows that the strong El Niño has more influence on rainfall anomalies in the post-1980s. Since there are insufficient studies on the association between teleconnections and climate variables, especially vertical mixing ratio, our findings can benefit prediction development for teleconnection-induced regional climate anomalies for extreme events and water management preparations in northern and northeastern Thailand.
HIGHLIGHTS
ENSO events have significant correlations in northern compared with northeastern Thailand.
El Niño, positive IOD, and climate variables have significant negative correlations in the dry season in northern Thailand.
La Niña, negative IOD, and climate variables have a less significant negative correlation in northern and northeastern Thailand.
MJO has significant correlations with rainfall and temperature.
INTRODUCTION
Teleconnection patterns are defined as recurring and persistent large-scale patterns of pressure and circulation anomalies that can influence long-distance regional weather patterns far from the source through wind circulations. The implication of weather patterns varies with regions and seasons (NOAA 2020; Clarke et al. 2022; Zavadoff & Arcodia 2022). The well-known teleconnections such as El Niño–Southern Oscillation (ENSO), Indian Ocean Dipole (IOD), and Madden–Julian Oscillation (MJO) are mostly sea surface temperature (SST) anomalies in the tropical Pacific, Indian Ocean, and Tropical areas (Feldstein & Franzke 2017). These events can influence significant global effects on weather variations, especially the precipitation distribution (Rodríguez-Fonseca et al. 2016; Cash et al. 2017; Aamir & Hassan 2020; Williams et al. 2023). Furthermore, some worldwide extreme weather events are found to be associated with teleconnection phenomena (Räsänen & Kummu 2013; Czymzik et al. 2016; Boers et al. 2019; Ciemer et al. 2020; Mishra et al. 2022; Liu et al. 2023). Therefore, the relationship between teleconnections and various climate variables should be investigated to improve the reliability of weather prediction systems.
The relationship between teleconnections, precipitation, and surface temperature has been examined by many studies, which mostly confirm the linkage of the negative teleconnection phase with lower-than-usual rainfall and the positive teleconnection phase with higher-than-usual rainfall (Phillips et al. 2012; Irwandi et al. 2018; Sreekala et al. 2018; Li et al. 2021). Aamir et al. (2022) ascertained that in Pakistan, the influence of teleconnection indices on monthly precipitation patterns is weak in winter months and moderate to strong in monsoon month. Moreover, Lüdecke et al. (2021) studied the influence of potential teleconnections with continent-wide African rainfall by 11-month time lag correlation to improve short- to mid-term rainfall prognoses. Lee et al. (2023) found that the United States precipitation responds to ENSO events with a high significance level of an opposite tendency. In addition, the likelihood that heat waves and extreme precipitation can occur in two consecutive seasons is higher during El Niño than La Niña events (Mishra et al. 2022). Even though the concern of global temperature rise from anthropogenic climate change might increase the teleconnection events, there is overall low confidence about the future changes of magnitude, frequency, and spatial distribution of these patterns (IPCC 2021).
Thailand's weather patterns are influenced by prevailing monsoons, namely the Southwest Monsoon, which brings abundant warm moist airmass from the Indian Ocean resulting in the rainy season from mid-May to mid-October (World Bank Group 2021). The Northeast Monsoon brings cold, dry airmass from mainland China to Thailand during the dry season in upper Thailand, namely, central, northern, and northeastern. However, this monsoon brings moisture from the Thai Gulf to southern Thailand resulting in abundant rainfall from mid-October to mid-February (Thai Meteorological Department 2015). The strong link between ENSO and monsoons in Thailand has been found since 1980 and was linked to several extreme events (Singhrattna et al. 2005). For instance, from 1997 to 1998, El Niño year, the observed rainfall amounts data from the Thai Meteorological Department (TMD) indicated below-normal rainfall by 5.47% and 14.46%, respectively, whereas in La Niña year 1999, the rainfall amount in Thailand was above normal by 36.59% (Ueangsawat 2013; Promchote et al. 2016).
Several studies have attempted to find the relationship between major teleconnections and extreme events in Thailand for better weather predictions (Sahu et al. 2010; Wikarmpapraharn & Kositsakulchai 2010; Muangsong et al. 2014; Puttrawutichai et al. 2017). Moreover, the studies showed that rainfall associated with ENSO, IOD, and MJO results in different magnitudes and timing for correlation values in different parts of Thailand. However, some studies revealed that the significant relationship between ENSO, IOD, and rainfall patterns could not be concluded (Nounmusig 2018; Bridhikitti 2019). Besides, vertical humidity and temperature determine the probability of magnitude and intensity of rainfall, which can lead to droughts or heavy rainfall. Nevertheless, the study of the relationship between teleconnections and vertical humidity in Thailand is still insufficient. Trakolkul et al. (2022) investigated the moderate to severe correlation between ENSO and precipitable water vapor (PWV) from 11 stations during 2007–2016. La Niña events were highly correlated with PWV, especially in Thailand's severe flooding in 2011. Furthermore, the PWV declined at the end of 2011 and consequently caused flooding and severe storms.
The percentage of the agricultural area in northern and northeastern Thailand is 23.6% and 46.8% (NSO 2013). However, northern and northeastern Thailand is susceptible to extreme weather, such as floods and droughts, adversely affecting rainfed and irrigated agricultural systems (UNDP 2023). Moreover, 80% of Thailand's severe drought-prone areas occur in the north and northeast regions (World Bank Group 2021). Agriculture is strongly affected by weather and climate variability (Hoogenboom 2000). The crop yield is sensitive to teleconnection patterns; therefore advance notice to farmers for adaptation and mitigation of any potential losses can be provided with proper analysis of the teleconnection influences (Gonsamu & Chen 2015; Ceglar et al. 2017). UNICEF (2023) reported that Thai children are at high risk from climate change and environmental degradation such as extreme heat, cold, floods, and droughts, especially in northeast and southern Thailand. In addition, Chiang Mai and Chiang Rai provinces in northern Thailand were ranked top ten in floods and cold risks. Furthermore, northern and northeastern areas may face riverine flooding during the rainy season (Lim & Boochabun 2012) and flash floods in mountainous areas or the foothills (Thai Meteorological Department 2015).
This study aims to determine the relationship patterns between major teleconnections and important climate variables that are crucial to extreme weather predictions. Furthermore, the implication of ENSO spatially and temporally was also investigated in the north and northeastern Thailand. The findings from this study could enhance our knowledge about the teleconnection implication of weather variation in northern and northeastern Thailand. In addition, the improvement of extreme weather prediction can be beneficial for disaster preparation and water resource management in northern and northeastern Thailand, which are vulnerable to various climatic disasters.
MATERIALS AND METHODS
The approach for relationship pattern establishment between major teleconnections and climate variables is based on statistical lag correlation to measure the paired variables’ association strength and direction. Subsequently, the statistical significance is confirmed with the t-test method. The data and methodology used in the analysis are described in the following sub-sections.
Study area
Northeastern Thailand's location is at latitude 14.13° to 18.45°N and longitude 100.9°E to 105.61°E (Figure 1). The dominant topography is a high plain with high mountain borders. This area consists of 20 provinces covering around 170,000 km2. The average annual rainfall is 1,480 mm, with approximately 88.7% of rainfall occurring in the rainy season. The annual average maximum temperature is 32 °C, and the highest average maximum temperature month is April (35.9 °C). The annual average minimum temperature is 21.6 °C, and the lowest average maximum temperature month is January (16.2 °C).
The main source of agricultural water in north and northeastern Thailand is monsoon rainfall. During the rainy season, a ‘dry spell’ may occur around the end of June to mid-July, resulting in low rainfall (Thai Meteorological Department 2015).
Data collection
The relationship patterns between teleconnections and climate variables were established using the selected teleconnection indices and main climate variables. Detail about the data collected is explained in Table 1 and subsequent sections.
No. . | Data . | Resolution . | Period . | Source . |
---|---|---|---|---|
I. | Teleconnection indices | |||
ONI (ENSO) | 3-MRM | 1950–2021 | NOAA | |
DMI (IOD) | Monthly | 1870–2021 | NOAA | |
RMM (MJO) | Daily | 1974–2021 | BOM | |
II. | Surface climate variables | |||
Rainfall | Point/Daily | 1951–2021 | TMD, RID | |
Maximum temperature | ||||
Minimum temperature | Point/Daily | 1951–2021 | TMD | |
III. | Vertical climate variables | |||
Upper-air data | Point/Daily | 1973–2021 | UW (TMD) | |
Vertical temperature | ||||
Vertical dew point temperature |
No. . | Data . | Resolution . | Period . | Source . |
---|---|---|---|---|
I. | Teleconnection indices | |||
ONI (ENSO) | 3-MRM | 1950–2021 | NOAA | |
DMI (IOD) | Monthly | 1870–2021 | NOAA | |
RMM (MJO) | Daily | 1974–2021 | BOM | |
II. | Surface climate variables | |||
Rainfall | Point/Daily | 1951–2021 | TMD, RID | |
Maximum temperature | ||||
Minimum temperature | Point/Daily | 1951–2021 | TMD | |
III. | Vertical climate variables | |||
Upper-air data | Point/Daily | 1973–2021 | UW (TMD) | |
Vertical temperature | ||||
Vertical dew point temperature |
Note: ONI, Oceanic Niño Index; ENSO, El Niño–Southern Oscillation; DMI, Dipole Mode Index; IOD, Indian Ocean Dipole; RMM, Real-time Multivariate MJO Index; MJO, Madden–Julian Oscillation; NOAA, National Oceanic and Atmospheric Administration; BOM, Bureau of Meteorology; TMD, Thai Meteorological Department; RID, Royal Irrigation Department; UW, University of Wyoming; 3-MRM, three-month running mean.
Surface climate variables
Observed daily rainfall, and the maximum and minimum surface temperature during 1951–2021 are acquired from TMD and Royal Irrigation Department (RID) weather stations. A total of 21 TMD weather stations distributed with daily data ≥ 30 years and missing data ≤ 10% across the study region were selected. The nearby station method filled in the missing data using the daily data from RID stations. The detail of paired TMD and RID stations are depicted in the Supplementary Material, Table S2.
Vertical climate variables
The observed daily data at Coordinated Universal Time (UTC) 0000Z during 1973–2021 were acquired from two TMD upper-air measurement stations, namely Chiangmai station (VTCC) in northern and Ubon Ratchathani station (VTUU) in northeastern Thailand. Moreover, the daily vertical dew point temperature and temperature at the surface, 850, 700, 500, and 300 hPa, which are standard levels, were used for the water vapor mixing ratio for representing vertical humidity and temperature calculations (MetOffice 2011). The upper-air measurement data used in this study were retrieved from the University of Wyoming website (https://weather.uwyo.edu/upperair/).
Teleconnection indices
Teleconnection indices are the ocean–atmosphere interaction events in one place that can influence weather patterns in a distant part of the world (Zavadoff & Arcodia 2022). The teleconnections selected for analysis in this study are ENSO, IOD, and MJO. For ENSO, the ONI which is NOAA's primary indicator for monitoring ENSO states was used; the ONI values are positive (negative) during El Niño (La Niña) phases (NOAA 2020). The second teleconnection index is the Dipole Mode Index (DMI), commonly used to identify IOD. IOD events may start in May or June, mature during August and October, and rapidly decay until around January (Bureau of Meteorology 2021a). The Real-time Multivariate MJO Index (RMM MJO) is the index for identifying MJO status. Overall, eight phases of MJO also affect opposite convective (suppressive) phases of the dipole (NOAA 2020).
ONI in 3-mrm values was used to find the relationship between ENSO, IOD, and climate variables in each month. Moreover, monthly DMI (IOD) was also used to find the association with monthly climate variables, so the association of both teleconnections to monthly climate variables could be compared. However, the MJO cycle is shorter than ENSO and IOD, therefore the relationship between MJO and climate variables was calculated weekly.
Overall methodology
The relationship association between teleconnections and climate variables and ENSO influence on spatial and temporal rainfall in northern and northeastern Thailand were assessed using different statistics. Since teleconnections might influence climate variables on different timescales, based on previous studies (Sattar et al. 2021; Wijeratne et al. 2022), the methods in this study are separated into two sections. Firstly, the lag correlation was applied to investigate teleconnections' influence on monthly and weekly climate variables. Secondly, the warm and cold phases of ENSO events' influence on annual rainfall were visualized in spatial and temporal maps.
Teleconnections vs climate variable association
The daily rainfall from weather stations was summarized into weekly and monthly data – subsequently, the average between northern and northeastern stations. Maximum and minimum temperatures were averaged into monthly and weekly data. The 95th percentile values for monthly rainfall, maximum temperature, and the fifth percentile for monthly minimum temperature were calculated to assess extreme conditions. Subsequently, the monthly and weekly climate variables were used to perform lag correlation and t-tests with negative and positive phases of teleconnection indices. The 3-mrm ONI (ENSO) performed lag correlation with monthly climate variables, and the first or zero lag starts with the first month of ONI; for instance, the ONI index in DJF month will correlate lag 0 with December rainfall, lag 1 with January rainfall and so on until lag 11 follows (Ueangsawat 2013). Furthermore, IOD resolution is monthly, therefore, the lag correlation was performed with monthly climate variables from 0 to 11 lags as well. Moreover, MJO vs climate variables were correlated by weekly lag correlation.
In this study, the correlation coefficient and t-test calculation were done using the Real Statistics Resource Pack for Excel free software (Zaiontz 2022). The positive and negative values of teleconnection indices and climate variable values can affect the direction of correlation. Therefore, the interpretation of correlation directions of the results between positive (negative) indices and positive (negative) variables is summarized in the Supplementary Material, Table S3.
Multiple testing
The multiple testing problem can occur in collective significance testing for determining significant patterns or an overall effect. This problem can lead to spurious significance results (Cortés et al. 2020). In this study, the statistical t-test was performed with an 11-lag correlation for each teleconnection index and climate variable, therefore our results may suffer from false-positive errors. The correction method such as the familywise error rate of Bonferroni can increase false-negative errors, due to very low adjusted p-values. The objective of this study is to explore the connection between teleconnections and regional climate variables in north and northeastern Thailand, therefore the correction of p-values is not employed. Hence, detailed confirmation of the present results may be required in future research (Lüdecke et al. 2021).
The spatiotemporal influence of ENSO on annual rainfall
The spatiotemporal influence of ENSO on north and northeast Thailand was studied using the Inverse Distance Weighted (IDW) method to construct rainfall anomaly maps from 1951–1980 and 1981–2021 (Chen & Liu 2012). Singhrattna et al. (2005) found a linkage between ENSO and Thailand monsoons was more closely related after 1980, which is contrary to the ENSO–Indian monsoon relationship. The ENSO indices are measured from SST anomalies, therefore the global temperature increase may affect the intensities of ENSO events. Hence, anomaly rainfall maps were constructed for 1951–1980 and 1981–2021 to investigate the temporal intensity between ENSO and rainfall. In addition, the influence of ENSO was considered by onset or ENSO start year and decay or ENSO decline year. Furthermore, the ENSO year intensities were classified into three groups based on the magnitude values of the ONI index. The threshold of each intensity is Strong ≥ 1.5, Moderate 1.0–1.4, and Weak 0.5–0.9; additional information on the ENSO state year is provided in the Supplementary Material, Table S4 (Golden Gate Weather Services 2022).
To investigate the implications of ENSO on different elevations, the weather stations were classified into two groups based on location heights. The threshold 200 m mean-sea-level (MSL) was chosen based on the height of the Khorat plateau, which covers most of northeastern Thailand. The weather stations at 200 m MSL or higher represent the highlands and mountainous areas, and those lower than 200 m MSL represent plain areas. Subsequently, the rainfall anomalies of ENSO events were visualized as contour maps.
RESULTS AND DISCUSSIONS
. | DJF . | JFM . | FMA . | MAM . | AMJ . | MJJ . | JJA . | JAS . | ASO . | SON . | OND . | NDJ . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
El Niño vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | − 0.49(4) | − 0.54(3) | − 0.52(1) | − 0.43(1) | − 0.48(1) | − 0.28(1) | − 0.73(0) | − 0.35(2) | 0.36(10) | − 0.27(7) | − 0.43(11) | − 0.46(5) |
R_95 | − 0.47(4) | − 0.44(3) | − 0.64(6) | 0.40(4) | − 0.29(9) | − 0.36(1) | − 0.71(0) | − 0.47(2) | − 0.30(6) | − 0.43(11) | − 0.50(11) | − 0.55(10) |
Northeast | ||||||||||||
R_M | − 0.41(3) | 0.42(7) | 0.49(4) | − 0.75(0) | − 0.44(1) | 0.37(9) | 0.36(8) | − 0.41(2) | 0.42(10) | 0.53(1) | 0.42(0) | − 0.42(10) |
R_95 | 0.46(6) | − 0.50(3) | 0.63(4) | − 0.75(0) | − 0.57(5) | 0.47(9) | 0.39(4) | 0.38(3) | 0.52(10) | 0.51(1) | 0.48(0) | − 0.45(10) |
La Niña vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | − 0.39(9) | − 0.28(3) | − 0.60(8) | 0.65(4) | 0.62(5) | − 0.60(3) | 0.43(3) | − 0.33(5) | − 0.41(8) | − 0.48(7) | 0.42(8) | 0.38(7) |
R_95 | − 0.34(5) | 0.40(7) | 0.49(7) | 0.43(4) | 0.66(5) | − 0.51(3) | 0.60(3) | 0.40(2) | 0.42(1) | 0.54(0) | 0.32(5) | 0.33(7) |
Northeast | ||||||||||||
R_M | − 0.31(4) | − 0.36(3) | − 0.33(4) | − 0.37(11) | 0.49(5) | 0.33(4) | 0.44(2) | 0.33(8) | − 0.36(11) | 0.55(9) | 0.33(8) | − 0.41(5) |
R_95 | − 0.27(7) | − 0.32(6) | − 0.32(3) | − 0.36(11) | 0.45(5) | − 0.32(8) | 0.37(2) | − 0.45(5) | − 0.22(11) | 0.49(9) | 0.26(1) | − 0.31(5) |
. | DJF . | JFM . | FMA . | MAM . | AMJ . | MJJ . | JJA . | JAS . | ASO . | SON . | OND . | NDJ . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
El Niño vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | − 0.49(4) | − 0.54(3) | − 0.52(1) | − 0.43(1) | − 0.48(1) | − 0.28(1) | − 0.73(0) | − 0.35(2) | 0.36(10) | − 0.27(7) | − 0.43(11) | − 0.46(5) |
R_95 | − 0.47(4) | − 0.44(3) | − 0.64(6) | 0.40(4) | − 0.29(9) | − 0.36(1) | − 0.71(0) | − 0.47(2) | − 0.30(6) | − 0.43(11) | − 0.50(11) | − 0.55(10) |
Northeast | ||||||||||||
R_M | − 0.41(3) | 0.42(7) | 0.49(4) | − 0.75(0) | − 0.44(1) | 0.37(9) | 0.36(8) | − 0.41(2) | 0.42(10) | 0.53(1) | 0.42(0) | − 0.42(10) |
R_95 | 0.46(6) | − 0.50(3) | 0.63(4) | − 0.75(0) | − 0.57(5) | 0.47(9) | 0.39(4) | 0.38(3) | 0.52(10) | 0.51(1) | 0.48(0) | − 0.45(10) |
La Niña vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | − 0.39(9) | − 0.28(3) | − 0.60(8) | 0.65(4) | 0.62(5) | − 0.60(3) | 0.43(3) | − 0.33(5) | − 0.41(8) | − 0.48(7) | 0.42(8) | 0.38(7) |
R_95 | − 0.34(5) | 0.40(7) | 0.49(7) | 0.43(4) | 0.66(5) | − 0.51(3) | 0.60(3) | 0.40(2) | 0.42(1) | 0.54(0) | 0.32(5) | 0.33(7) |
Northeast | ||||||||||||
R_M | − 0.31(4) | − 0.36(3) | − 0.33(4) | − 0.37(11) | 0.49(5) | 0.33(4) | 0.44(2) | 0.33(8) | − 0.36(11) | 0.55(9) | 0.33(8) | − 0.41(5) |
R_95 | − 0.27(7) | − 0.32(6) | − 0.32(3) | − 0.36(11) | 0.45(5) | − 0.32(8) | 0.37(2) | − 0.45(5) | − 0.22(11) | 0.49(9) | 0.26(1) | − 0.31(5) |
Note: Numbers in () denote lag months, bold values indicate p-value < 0.05.
R_M indicates average monthly rainfall, R_95 indicates 95th percentile monthly rainfall. El Niño ONI values are positive (plus values) and La Niña ONI values are negative (minus values).
. | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Positive IOD vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | 0.80(2) | − 0.83(4) | 0.95(5) | − 1.00(6) | 0.81(2) | − 0.68(7) | − 0.75(8) | − 0.71(8) | − 0.55(11) | − 0.71(11) | ||
R_95 | 0.80(3) | − 0.77(4) | 0.95(7) | 0.80(2) | − 0.68(7) | − 0.77(7) | − 0.49(8) | − 0.47(8) | − 0.47(9) | − 0.70(11) | ||
Northeast | ||||||||||||
R_M | 1.00(10) | − 0.54(8) | 0.95(5) | 1.00(5) | − 0.74(7) | − 0.71(7) | − 0.63(8) | − 0.65(2) | − 0.66(6) | − 0.48(1) | ||
R_95 | 1.00(10) | 0.83(5) | 0.95(5) | − 1.00(8) | − 0.84(7) | − 0.77(7) | − 0.72(8) | − 0.62(2) | − 0.79(6) | − 0.55(7) | ||
Negative IOD vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | 0.79(6) | 0.80(0) | − 0.89(4) | 0.53(5) | − 0.58(1) | − 0.53(7) | 0.81(8) | 0.53(11) | − 0.41(4) | 0.94(2) | ||
R_95 | 0.75(0) | 0.80(1) | − 0.71(4) | 0.51(5) | 0.57(5) | 0.49(5) | 0.52(8) | 0.41(9) | − 0.58(4) | 0.94(2) | ||
Northeast | ||||||||||||
R_M | 0.93(5) | 0.70(2) | − 0.77(4) | − 0.55(10) | 0.97(5) | − 0.27(9) | 0.72(8) | 0.46(9) | − 0.42(0) | 0.89(2) | ||
R_95 | 0.86(5) | 0.90(2) | − 0.89(4) | − 0.57(10) | 0.97(5) | 0.32(6) | 0.45(8) | 0.52(9) | − 0.49(0) | 0.84(2) |
. | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Positive IOD vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | 0.80(2) | − 0.83(4) | 0.95(5) | − 1.00(6) | 0.81(2) | − 0.68(7) | − 0.75(8) | − 0.71(8) | − 0.55(11) | − 0.71(11) | ||
R_95 | 0.80(3) | − 0.77(4) | 0.95(7) | 0.80(2) | − 0.68(7) | − 0.77(7) | − 0.49(8) | − 0.47(8) | − 0.47(9) | − 0.70(11) | ||
Northeast | ||||||||||||
R_M | 1.00(10) | − 0.54(8) | 0.95(5) | 1.00(5) | − 0.74(7) | − 0.71(7) | − 0.63(8) | − 0.65(2) | − 0.66(6) | − 0.48(1) | ||
R_95 | 1.00(10) | 0.83(5) | 0.95(5) | − 1.00(8) | − 0.84(7) | − 0.77(7) | − 0.72(8) | − 0.62(2) | − 0.79(6) | − 0.55(7) | ||
Negative IOD vs Monthly Rainfall | ||||||||||||
North | ||||||||||||
R_M | 0.79(6) | 0.80(0) | − 0.89(4) | 0.53(5) | − 0.58(1) | − 0.53(7) | 0.81(8) | 0.53(11) | − 0.41(4) | 0.94(2) | ||
R_95 | 0.75(0) | 0.80(1) | − 0.71(4) | 0.51(5) | 0.57(5) | 0.49(5) | 0.52(8) | 0.41(9) | − 0.58(4) | 0.94(2) | ||
Northeast | ||||||||||||
R_M | 0.93(5) | 0.70(2) | − 0.77(4) | − 0.55(10) | 0.97(5) | − 0.27(9) | 0.72(8) | 0.46(9) | − 0.42(0) | 0.89(2) | ||
R_95 | 0.86(5) | 0.90(2) | − 0.89(4) | − 0.57(10) | 0.97(5) | 0.32(6) | 0.45(8) | 0.52(9) | − 0.49(0) | 0.84(2) |
Note: Numbers in () denote lag months, bold values indicate p-value < 0.05.
R_M indicates average monthly rainfall, R_95 indicates 95th percentile monthly rainfall. Positive IOD values are positive (plus values). Negative IOD values are negative (minus values).
. | DJF . | JFM . | FMA . | MAM . | AMJ . | MJJ . | JJA . | JAS . | ASO . | SON . | OND . | NDJ . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
El Niño vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.68(4) | 0.66(3) | 0.57(6) | 0.61(5) | 0.33(2) | 0.41(5) | 0.59(9) | 0.48(8) | − 0.39(10) | 0.29(7) | 0.48(6) | 0.59(5) |
T_95 | 0.56(4) | 0.57(3) | 0.63(1) | 0.70(0) | 0.57(4) | 0.42(9) | 0.51(9) | 0.42(8) | 0.40(0) | 0.25(1) | 0.34(6) | 0.49(5) |
Northeast | ||||||||||||
T_Max | 0.53(4) | 0.55(2) | 0.50(2) | 0.73(1) | − 0.37(6) | 0.43(10) | 0.51(9) | 0.50(4) | 0.40(5) | 0.28(10) | 0.44(6) | 0.49(5) |
T_95 | 0.47(3) | 0.55(2) | 0.62(3) | 0.62(0) | − 0.28(7) | 0.59(10) | 0.55(9) | 0.46(8) | 0.41(5) | 0.39(6) | 0.48(5) | 0.47(4) |
La Niña vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.44(5) | 0.42(1) | 0.46(1) | 0.46(9) | 0.46(2) | 0.51(7) | 0.48(10) | 0.49(9) | 0.61(8) | − 0.53(10) | 0.34(11) | 0.38(6) |
T_95 | 0.36(4) | 0.41(3) | 0.43(9) | 0.52(8) | 0.46(7) | − 0.42(9) | 0.41(10) | 0.42(11) | 0.58(8) | 0.57(7) | 0.37(6) | 0.40(5) |
Northeast | ||||||||||||
T_Max | 0.48(2) | 0.73(1) | 0.58(0) | 0.45(3) | 0.56(2) | 0.49(7) | 0.48(10) | 0.50(3) | 0.43(8) | − 0.39(9) | 0.54(0) | 0.43(3) |
T_95 | 0.31(5) | 0.46(3) | 0.42(4) | 0.44(8) | 0.41(2) | 0.42(3) | − 0.60(3) | 0.44(11) | 0.46(9) | 0.52(7) | 0.44(7) | 0.42(6) |
. | DJF . | JFM . | FMA . | MAM . | AMJ . | MJJ . | JJA . | JAS . | ASO . | SON . | OND . | NDJ . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
El Niño vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.68(4) | 0.66(3) | 0.57(6) | 0.61(5) | 0.33(2) | 0.41(5) | 0.59(9) | 0.48(8) | − 0.39(10) | 0.29(7) | 0.48(6) | 0.59(5) |
T_95 | 0.56(4) | 0.57(3) | 0.63(1) | 0.70(0) | 0.57(4) | 0.42(9) | 0.51(9) | 0.42(8) | 0.40(0) | 0.25(1) | 0.34(6) | 0.49(5) |
Northeast | ||||||||||||
T_Max | 0.53(4) | 0.55(2) | 0.50(2) | 0.73(1) | − 0.37(6) | 0.43(10) | 0.51(9) | 0.50(4) | 0.40(5) | 0.28(10) | 0.44(6) | 0.49(5) |
T_95 | 0.47(3) | 0.55(2) | 0.62(3) | 0.62(0) | − 0.28(7) | 0.59(10) | 0.55(9) | 0.46(8) | 0.41(5) | 0.39(6) | 0.48(5) | 0.47(4) |
La Niña vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.44(5) | 0.42(1) | 0.46(1) | 0.46(9) | 0.46(2) | 0.51(7) | 0.48(10) | 0.49(9) | 0.61(8) | − 0.53(10) | 0.34(11) | 0.38(6) |
T_95 | 0.36(4) | 0.41(3) | 0.43(9) | 0.52(8) | 0.46(7) | − 0.42(9) | 0.41(10) | 0.42(11) | 0.58(8) | 0.57(7) | 0.37(6) | 0.40(5) |
Northeast | ||||||||||||
T_Max | 0.48(2) | 0.73(1) | 0.58(0) | 0.45(3) | 0.56(2) | 0.49(7) | 0.48(10) | 0.50(3) | 0.43(8) | − 0.39(9) | 0.54(0) | 0.43(3) |
T_95 | 0.31(5) | 0.46(3) | 0.42(4) | 0.44(8) | 0.41(2) | 0.42(3) | − 0.60(3) | 0.44(11) | 0.46(9) | 0.52(7) | 0.44(7) | 0.42(6) |
Note: Numbers in () denote lag months, bold values indicate p-value < 0.05.
T_Max indicates average monthly maximum surface temperature, T_95 indicates 95th percentile monthly maximum surface temperature. El Niño ONI values are positive (plus values). La Niña ONI values are negative (minus values).
. | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Positive IOD vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.80(5) | 0.83(4) | 0.95(4) | 0.80(7) | 0.88(8) | − 0.68(1) | 0.69(4) | 0.78(8) | 0.78(8) | 0.64(1) | ||
T_95 | 1.00(7) | 0.71(6) | 0.95(7) | 1.00(9) | 0.62(8) | − 0.88(1) | − 0.59(0) | 0.84(8) | 0.80(9) | 0.69(3) | ||
Northeast | ||||||||||||
T_Max | 0.80(4) | 0.77(4) | 0.95(4) | 0.80(7) | 0.81(8) | − 0.68(2) | 0.83(4) | 0.77(8) | 0.82(8) | 0.69(4) | ||
T_95 | 0.80(3) | 0.77(6) | 0.95(4) | 1.00(9) | 0.76(8) | − 0.64(1) | 0.70(4) | 0.89(11) | 0.73(8) | 0.72(3) | ||
Negative IOD vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.96(11) | − 1.00(5) | 0.77(5) | 0.53(3) | 0.65(4) | 0.37(2) | − 0.91(8) | − 0.54(10) | 0.32(11) | 0.66(1) | ||
T_95 | 0.86(10) | 0.90(8) | − 0.77(11) | 0.50(9) | − 0.82(1) | − 0.65(6) | − 0.71(8) | 0.38(0) | 0.41(11) | 0.66(5) | ||
Northeast | ||||||||||||
T_Max | 0.82(11) | 1.00(8) | 0.94(4) | − 0.45(7) | 0.72(9) | 0.41(8) | − 0.71(8) | 0.52(2) | 0.45(2) | 0.77(1) | ||
T_95 | 0.64(11) | 1.00(8) | 0.71(4) | − 0.38(10) | − 0.60(0) | 0.38(8) | − 0.61(8) | − 0.37(10) | 0.37(5) | − 0.77(9) |
. | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Positive IOD vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.80(5) | 0.83(4) | 0.95(4) | 0.80(7) | 0.88(8) | − 0.68(1) | 0.69(4) | 0.78(8) | 0.78(8) | 0.64(1) | ||
T_95 | 1.00(7) | 0.71(6) | 0.95(7) | 1.00(9) | 0.62(8) | − 0.88(1) | − 0.59(0) | 0.84(8) | 0.80(9) | 0.69(3) | ||
Northeast | ||||||||||||
T_Max | 0.80(4) | 0.77(4) | 0.95(4) | 0.80(7) | 0.81(8) | − 0.68(2) | 0.83(4) | 0.77(8) | 0.82(8) | 0.69(4) | ||
T_95 | 0.80(3) | 0.77(6) | 0.95(4) | 1.00(9) | 0.76(8) | − 0.64(1) | 0.70(4) | 0.89(11) | 0.73(8) | 0.72(3) | ||
Negative IOD vs maximum surface temperature | ||||||||||||
North | ||||||||||||
T_Max | 0.96(11) | − 1.00(5) | 0.77(5) | 0.53(3) | 0.65(4) | 0.37(2) | − 0.91(8) | − 0.54(10) | 0.32(11) | 0.66(1) | ||
T_95 | 0.86(10) | 0.90(8) | − 0.77(11) | 0.50(9) | − 0.82(1) | − 0.65(6) | − 0.71(8) | 0.38(0) | 0.41(11) | 0.66(5) | ||
Northeast | ||||||||||||
T_Max | 0.82(11) | 1.00(8) | 0.94(4) | − 0.45(7) | 0.72(9) | 0.41(8) | − 0.71(8) | 0.52(2) | 0.45(2) | 0.77(1) | ||
T_95 | 0.64(11) | 1.00(8) | 0.71(4) | − 0.38(10) | − 0.60(0) | 0.38(8) | − 0.61(8) | − 0.37(10) | 0.37(5) | − 0.77(9) |
Note: Numbers in () denote lag months, bold values indicate p-value < 0.05.
T_Max indicates average monthly temperature, T_95 indicates 95th percentile monthly temperature. Positive IOD values are positive (plus values) and negative IOD values are negative (minus values).
. | DJF . | JFM . | FMA . | MAM . | AMJ . | MJJ . | JJA . | JAS . | ASO . | SON . | OND . | NDJ . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
El Niño vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | − 0.26(11) | 0.34(4) | 0.37(3) | 0.36(1) | − 0.32(8) | 0.52(10) | 0.47(5) | 0.56(4) | 0.44(3) | 0.39(2) | 0.37(1) | 0.34(1) |
T_05 | 0.44(0) | 0.36(4) | 0.42(11) | − 0.45(8) | 0.57(5) | 0.60(7) | 0.63(5) | 0.76(4) | 0.64(3) | 0.56(2) | 0.53(1) | 0.47(1) |
Northeast | ||||||||||||
T_Min | 0.28(4) | 0.31(3) | − 0.44(0) | − 0.41(8) | − 0.57(10) | − 0.27(4) | 0.39(9) | 0.44(4) | − 0.28(1) | − 0.24(11) | 0.28(2) | 0.32(1) |
T_05 | 0.52(4) | 0.45(3) | − 0.24(0) | 0.50(1) | − 0.70(10) | 0.35(11) | 0.64(5) | 0.72(4) | 0.41(3) | 0.30(7) | 0.46(6) | 0.47(5) |
La Niña vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | 0.26(5) | 0.31(4) | 0.26(3) | 0.41(2) | − 0.41(9) | 0.37(7) | 0.35(6) | 0.41(10) | 0.31(9) | 0.39(4) | 0.41(2) | 0.30(1) |
T_05 | 0.30(5) | 0.27(4) | 0.24(9) | 0.45(9) | − 0.55(0) | − 0.32(4) | 0.40(6) | 0.49(3) | 0.42(2) | 0.43(1) | 0.53(0) | 0.41(1) |
Northeast | ||||||||||||
T_Min | 0.45(2) | 0.57(1) | 0.61(0) | 0.35(2) | 0.35(1) | 0.31(7) | 0.30(9) | 0.37(8) | 0.23(9) | 0.31(3) | 0.40(1) | 0.47(3) |
T_05 | 0.42(2) | 0.55(1) | 0.58(0) | 0.33(2) | 0.40(1) | − 0.32(4) | 0.36(8) | 0.40(3) | 0.37(2) | 0.43(3) | 0.60(0) | 0.46(1) |
. | DJF . | JFM . | FMA . | MAM . | AMJ . | MJJ . | JJA . | JAS . | ASO . | SON . | OND . | NDJ . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
El Niño vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | − 0.26(11) | 0.34(4) | 0.37(3) | 0.36(1) | − 0.32(8) | 0.52(10) | 0.47(5) | 0.56(4) | 0.44(3) | 0.39(2) | 0.37(1) | 0.34(1) |
T_05 | 0.44(0) | 0.36(4) | 0.42(11) | − 0.45(8) | 0.57(5) | 0.60(7) | 0.63(5) | 0.76(4) | 0.64(3) | 0.56(2) | 0.53(1) | 0.47(1) |
Northeast | ||||||||||||
T_Min | 0.28(4) | 0.31(3) | − 0.44(0) | − 0.41(8) | − 0.57(10) | − 0.27(4) | 0.39(9) | 0.44(4) | − 0.28(1) | − 0.24(11) | 0.28(2) | 0.32(1) |
T_05 | 0.52(4) | 0.45(3) | − 0.24(0) | 0.50(1) | − 0.70(10) | 0.35(11) | 0.64(5) | 0.72(4) | 0.41(3) | 0.30(7) | 0.46(6) | 0.47(5) |
La Niña vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | 0.26(5) | 0.31(4) | 0.26(3) | 0.41(2) | − 0.41(9) | 0.37(7) | 0.35(6) | 0.41(10) | 0.31(9) | 0.39(4) | 0.41(2) | 0.30(1) |
T_05 | 0.30(5) | 0.27(4) | 0.24(9) | 0.45(9) | − 0.55(0) | − 0.32(4) | 0.40(6) | 0.49(3) | 0.42(2) | 0.43(1) | 0.53(0) | 0.41(1) |
Northeast | ||||||||||||
T_Min | 0.45(2) | 0.57(1) | 0.61(0) | 0.35(2) | 0.35(1) | 0.31(7) | 0.30(9) | 0.37(8) | 0.23(9) | 0.31(3) | 0.40(1) | 0.47(3) |
T_05 | 0.42(2) | 0.55(1) | 0.58(0) | 0.33(2) | 0.40(1) | − 0.32(4) | 0.36(8) | 0.40(3) | 0.37(2) | 0.43(3) | 0.60(0) | 0.46(1) |
Note: Numbers in () denote lag months, bold values indicate p-value < 0.05.
T_Min indicates average monthly minimum surface temperature, T_05 indicates fifth percentile monthly minimum surface temperature. El Niño ONI values are positive (plus values). La Niña ONI values are negative (minus values).
. | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Positive IOD vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | − 0.80(1) | 0.60(3) | 0.95(4) | 0.80(10) | 0.21(0) | − 0.65(0) | − 0.61(2) | 0.66(8) | 0.59(10) | 0.59(7) | ||
T_05 | − 0.80(1) | − 0.77(7) | 0.95(2) | − 0.60(11) | 0.57(1) | − 0.75(2) | − 0.59(7) | 0.65(8) | 0.63(6) | 0.58(7) | ||
Northeast | ||||||||||||
T_Min | 0.80(4) | − 0.60(10) | 0.74(11) | − 0.80(4) | 0.90(8) | − 0.52(11) | 0.69(4) | 0.68(8) | 0.53(9) | 0.90(4) | ||
T_05 | 1.00(8) | − 0.71(10) | 0.95(9) | − 0.60(4) | 0.88(9) | − 0.55(0) | 0.75(8) | 0.60(8) | 0.51(9) | 0.79(4) | ||
Negative IOD vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | 0.96(11) | 1.00(9) | 0.89(9) | 0.55(0) | 0.68(5) | 0.57(4) | − 0.42(10) | 0.63(2) | − 0.47(0) | 0.77(2) | ||
T_05 | 0.71(11) | − 0.80(4) | 0.77(9) | 0.69(11) | 0.77(4) | 0.59(8) | − 0.38(10) | 0.52(3) | − 0.64(0) | 0.77(1) | ||
Northeast | ||||||||||||
T_Min | 0.79(4) | 0.90(8) | 0.77(9) | 0.46(5) | 0.62(5) | 0.43(4) | 0.47(11) | 0.58(2) | − 0.62(4) | 0.77(5) | ||
T_05 | 0.79(6) | 0.70(0) | 0.37(0) | 0.52(11) | − 0.72(10) | 0.59(4) | 0.47(11) | 0.58(2) | − 0.54(0) | 0.83(1) |
. | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Positive IOD vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | − 0.80(1) | 0.60(3) | 0.95(4) | 0.80(10) | 0.21(0) | − 0.65(0) | − 0.61(2) | 0.66(8) | 0.59(10) | 0.59(7) | ||
T_05 | − 0.80(1) | − 0.77(7) | 0.95(2) | − 0.60(11) | 0.57(1) | − 0.75(2) | − 0.59(7) | 0.65(8) | 0.63(6) | 0.58(7) | ||
Northeast | ||||||||||||
T_Min | 0.80(4) | − 0.60(10) | 0.74(11) | − 0.80(4) | 0.90(8) | − 0.52(11) | 0.69(4) | 0.68(8) | 0.53(9) | 0.90(4) | ||
T_05 | 1.00(8) | − 0.71(10) | 0.95(9) | − 0.60(4) | 0.88(9) | − 0.55(0) | 0.75(8) | 0.60(8) | 0.51(9) | 0.79(4) | ||
Negative IOD vs minimum surface temperature | ||||||||||||
North | ||||||||||||
T_Min | 0.96(11) | 1.00(9) | 0.89(9) | 0.55(0) | 0.68(5) | 0.57(4) | − 0.42(10) | 0.63(2) | − 0.47(0) | 0.77(2) | ||
T_05 | 0.71(11) | − 0.80(4) | 0.77(9) | 0.69(11) | 0.77(4) | 0.59(8) | − 0.38(10) | 0.52(3) | − 0.64(0) | 0.77(1) | ||
Northeast | ||||||||||||
T_Min | 0.79(4) | 0.90(8) | 0.77(9) | 0.46(5) | 0.62(5) | 0.43(4) | 0.47(11) | 0.58(2) | − 0.62(4) | 0.77(5) | ||
T_05 | 0.79(6) | 0.70(0) | 0.37(0) | 0.52(11) | − 0.72(10) | 0.59(4) | 0.47(11) | 0.58(2) | − 0.54(0) | 0.83(1) |
Note: Numbers in () denote lag months, bold values indicate p-value < 0.05.
T_Min indicates average monthly minimum surface temperature and T_05 indicates fifth Percentile monthly minimum surface temperature. Positive IOD values are positive (plus values) and negative IOD values are negative (minus values).
Teleconnections vs rainfall
ENSO vs rainfall
The correlations between ENSO and monthly rainfall (R_M) and 95th percentile rainfall (R_95) in northern and northeastern Thailand were found moderately correlated. However, El Niño found more associations in northern than in northeast Thailand. In northern Thailand, El Niño occurred at JFM season was correlated at −0.54 with April R_M, which is similar to the findings of Ueangsawat (2013) of JFM SOI indices at 0.51 with May rainfall. In the northeast, El Niño occurred during the MAM season, and March R_M correlated at −0.75, and El Niño AMJ correlated with September R_95 at −0.57, indicating that El Niño may be associated with R_M and R_95 rainfall, during the dry season and the dry spell. However, northern and northeastern Thailand also found a positive correlation in some months.
There were not so many correlations between La Niña with R_M and R_95 in northern and northeastern Thailand. La Niña state is represented with negative values, therefore the negative direction of correlation between La Niña and rainfall means that when La Niña is stronger, rainfall can be higher than usual. In this study, only moderate correlations (−0.60) were found in La Niña FMA and October R_M, and La Niña MJJ and August R_M in northern Thailand as depicted in Table 2.
The above results of ENSO vs rainfall are similar to those of past studies in other Asian areas. Räsänen & Kummu (2013) found a moderate negative correlation between ENSO and rainfall in the Mekong basin, especially in northeast Thailand. Mishra et al. (2020) also found a moderate negative correlation between observed rainfall and ENSO variability in India. Moreover, the monthly lag times between La Niña events and rainfall in Korea and Japan were found at 4–5 lag months (Kawamura et al. 2005). In addition, extreme rainfall which caused devastating floods in Australia during 2010–2011 was found to have a strong link with La Niña events (Ummenhofer et al. 2015).
IOD vs rainfall
The correlation between positive IOD and rainfall in northern and northeastern Thailand is found to be negatively moderate, similar to El Niño. as depicted in Table 3. In northern Thailand, July–November p_IOD shows a moderate correlation with next year's R_M and R_95 in the dry season (February and April), early rainy season (May), rainy season (September), and late rainy season (October). Similarly, in the northeast, July–October p-IOD is associated with February, April, November, R_M, and February, April R_95, which may be associated with droughts in these areas. Cold phases or negative IOD (n-IOD) and R_M, R_95 also show moderate positive and negative associations. In contrast, Bridhikitti (2019) found no significant correlation between IOD and northern rainfall in May and April, however, IOD likely co-occurred with ENSO. In addition, the results of Ummenhofer et al. (2013) indicated that the co-occurrence between positive IOD and El Niño events could have worsened severe droughts in the Indo-Pacific region from 1877 to 2006.
MJO vs rainfall
Negligible values of correlation coefficient (ρ less than 0.2) were found between MJO and weekly rainfall. These results differ from the moderate −0.54 correlation coefficient values between MJO and May rainfall in northern Thailand (Bridhikitti 2019). Furthermore, Xavier et al. (2014) found a relationship between increased (decreased) extreme rainfall probability and convective (suppressive) MJO phases in Southeast Asia. Details of the correlation coefficient for MJO vs rainfall are depicted in the Supplementary Material, Table S7.
Teleconnections vs maximum surface temperature
ENSO vs maximum surface temperature
The positive correlations between El Niño, T_Max, and T_95 in northern and northeastern Thailand indicate greater warmth than usual in the dry season (April, March) and the rainy season (May and August) as depicted in Table 4. The increasing maximum surface temperature can intensify the evaporation process, whereas La Niña can be associated with daytime temperature reduction (Bureau of Meteorology 2021b). The positive direction of La Niña vs T_Max, T_95 shows that when La Niña is stronger, the T_Max and T_95 will be colder (see the details of correlation directions in the Supplementary Material, Table S3). A moderate positive correlation (0.4–0.6) was found between La Niña in April and May T_Max in northern Thailand. In northeastern Thailand, positive correlations were mostly found between La Niña and the dry season (February, April) and the rainy season (June, October) of T_Max.
IOD vs maximum surface temperature
The positive IOD (p_IOD) can influence higher-than-usual temperature, in contrast with negative IOD (n_IOD). The p_IOD are more correlated to T_Max and T_95 than n_IOD, as depicted in Table 5. In northern Thailand, the dry season (December), the rainy season (May), and the dry spell month (June) T_Max were positively correlated with p_IOD, whereas a negative correlation can be found between p_IOD and the rainy season (August) T_Max and T-95. The positive correlation in northeastern Thailand was also found between p_IOD and December, May, June T_Max. However, the n_IOD has both negative and positive correlations with T_Max and T_95 in northern and northeastern Thailand. The moderate to high negative correlation values were found between August n_IOD and April T_Max and T_95 in both northern and northeastern Thailand.
MJO vs maximum surface temperature
MJO and weekly maximum surface temperature also show a mostly positive correlation in both phases. However, the highest magnitude of the correlation is less than 0.2, which is negligible. Details of the correlation coefficient for MJO vs maximum surface temperature are depicted in the Supplementary Material, Table S8.
Teleconnections vs minimum surface temperature
ENSO vs minimum surface temperature
El Niño and La Niña have both moderate positive and negative correlations with T_Min and T_05 in northern and northeastern Thailand as depicted in Table 6. The positive correlation between El Niño and T_Min implies that when El Niño is stronger, T_Min is warmer. In contrast, a positive correlation between La Niña and T_Min implies that when La Niña is stronger, T_Min is colder. The positive moderate correlation between El Niño and T_Min, T_05 was found in northern Thailand more than in northeastern Thailand. In northern Thailand, a moderate positive correlation was found between El Niño and March, November T_Min. For T_05, a positive correlation was found between El Niño and September, November, and December T_05. For northeastern Thailand, a positive correlation was found between April and November T_05. The correlation between La Niña and October, December T_Min and T_05 in northern Thailand was also positive. In northeastern Thailand, a positive correlation was found between La Niña and February T_Min, and February, October, and December T_05.
IOD vs minimum surface temperature
The positive IOD (p_IOD) and negative IOD (n_IOD) correlate positively and negatively with T_Min and T_05 in northern and northeastern Thailand, similar to ENSO. A moderate positive correlation in northern Thailand was found between p_IOD and May, June, and August T_Min, and between p_IOD and April, May, and June T_05. In northeastern Thailand, the p_IOD was found moderately correlated with March, May, December T_Min, and March, April T_05. The n_IOD was found moderately correlated with May, November T_Min, and March, December T_05 in northern Thailand. In northeastern Thailand, the n_IOD was found positively correlated with November T_Min, T_05.
MJO vs minimum surface temperature
MJO and T_Min also show correlation coefficient values less than 0.2 but with mostly negative correlation. Details of correlation coefficient results for MJO vs minimum surface temperature are depicted in the Supplementary Material, Table S9.
ENSO vs vertical humidity
Vertical humidity is one of the crucial components of cloud formation. The surface to 700 hPa level is the level at which clouds form and grow, therefore the variation of vertical humidity at this level can affect cloud amounts and might affect rainfall variation. The correlation coefficient (CC) values between El Niño, monthly mixing ratio (MR), and 95th percentile monthly mixing ratio (MR_95) in northern Thailand are shown in Figure 3. Moderate negative correlations were found in March and April MR, and it coincides with previous results of negative correlation between El Niño and April R_M and El Niño and April T_Max positive correlations. The highest CC value −0.71 was found at the surface level of the 95th percentile monthly mixing ratio (MR_95) in March.
In northeastern Thailand, the correlation between El Niño and MR was mainly positive as depicted in Figure 4. The moderate negative correlation between El Niño and MR_95 was rarely found in northeastern Thailand, which contrasts with northern Thailand. The positive correlation between El Niño and MR implies that when El Niño is strong, MR also increases. However, not only MR amounts but also other variables should be considered for clouds and rainfall variation, such as temperature and wind patterns. Furthermore, mixing ratio data was calculated from only one point at the upper-air station, therefore the area and distance from the ocean may also be affected by the correlations.
The correlation results between La Niña and MR, MR_95 in northern Thailand, showed a more negative correlation than El Niño as depicted in Figure 5. At 850 hPa level, moderate negative correlations were found with March, April MR, and April MR_95. Trakolkul et al. (2022) indicated that La Niña events can induce water vapor increase in Thailand, especially as during the severe floods of 2011. However, the El Niño event has no relationship with water vapor, contrary to our results.
The correlation results between La Niña, MR, and MR_95 in northeastern Thailand are depicted in Figure 6. A negative correlation was found between La Niña and MR; MR_95 is negatively correlated between AMJ La Niña and April, June, October MR. In addition, a negative correlation between AMJ La Niña and April, October MR_95 was also presented.
ENSO vs vertical temperature
The correlation results between ENSO, monthly vertical temperature (VT), and 95th percentile monthly vertical temperature (VT_95) were found more correlated with La Niña than El Niño in northern and northeastern Thailand. The vertical temperatures at 300 and 500 hPa are lower than 0 °C, thus their values are negative as described in the Supplementary Material, Table S3. The results between El Niño, VT, and VT_95 in northern Thailand are depicted in Figure 7.
The negative correlation between El Niño and August VT at 300 and 500 hPa levels indicates an inverse relationship, which can imply that stronger El Niño can be associated with colder August VT. In contrast, a moderate positive correlation is also found between El Niño and April VT, which can imply that stronger El Niño can be associated with warmer April VT in the higher level.
A positive correlation between El Niño and Sfc April VT and February VT is presented in northeastern Thailand. In addition, El Niño and 700 hPa March VT also show a positive correlation as depicted in Figure 8. The positive correlation implies that when El Niño is stronger, VT is colder in the dry season. A negative correlation is also presented between El Niño and June VT_95 at the 850 hPa level, which implies colder June VT_95 when El Niño is strong.
La Niña events can be associated with the decreasing of vertical temperature. At the 300 and 500 hPa levels, the positive correlation direction indicates decreasing temperature when La Niña is stronger. Moreover, the significant results between La Niña, VT, and VT_95 were found to correlate during the rainy season in August, May, June, and September in northern Thailand as shown in Figure 9.
The positive correlation results between La Niña and VT, VT_95 in northeastern Thailand were found more significant at higher levels, which implies the inverse relationship between La Niña and higher-level VT in June. In contrast to northern Thailand, the significant correlations of northeastern Thailand were found mostly during June, December, and October as depicted in Figure 10.
Spatiotemporal influence of ENSO on average rainfall anomalies
Average annual rainfall anomaly maps depicted in Figure 12 are based on La Niña intensities in onset and decay years during pre- and post-1980. Despite this, the strong La Niña show more positive rainfall anomalies in the pre-1980 period than those of post-1980. The decay years of the strong La Niña show more positive rainfall anomalies in post-1980 in both northern and northeastern Thailand. In contrast, moderate La Niña shows higher positive rainfall anomalies in onset years and weak La Niña shows higher positive rainfall anomalies in both onset and decay years in post-1980. However, the decay years of moderate La Niña show higher positive rainfall anomalies in pre-1980.
The two periods of (1) 1951–1980 and (2) 1981–2021 were used to investigate the frequency and intensity of ENSO events as per the details in the Supplementary Material, Table S4. During both periods, there were a total of 26 El Niño events, period (1) 13 events which are equal to period (2). The strong El Niño in period (1) was three events and in period (2) was five. Moreover, in period (1) the moderate El Niño occurred three times, and four times for period (2). Therefore, the intensities of El Niño events were slightly increasing during period (2). In contrast, the frequency and intensity of La Niña events during period (2) were increasing. The La Niña events during period (1) were eight, whereas the La Niña events during period (2) were doubled to 16. Furthermore, the strong La Niña were two and five during periods (1) and (2), respectively. In addition, moderate La Niña events increased from two events in period (1) to four events in period (2).
ENSO impact on different elevations
CONCLUSION
To establish the relationship between major teleconnections and regional weather of northern and northeastern Thailand, the lag correlations between ENSO, IOD, MJO indices, and 21 TMD weather stations' rainfall, maximum, and minimum surface temperature data and two upper-air stations' vertical humidity and average temperature data were performed. Subsequently, the statistical significance was performed using a t-test at a 95% confidence level (p-value < 0.05). The results clearly showed that ENSO and IOD have a moderate relationship with climate variables. The El Niño or warm–dry phase of ENSO showed a negative significant relationship with monthly rainfall at the highest values of −0.73 and −0.75 in northern and northeastern Thailand, respectively. Furthermore, El Niño showed a significant negative relationship with 95th percentile monthly rainfall at the highest values of −0.71 and −0.75 in northern and northeastern Thailand, respectively. Moreover, La Niña depicted a more significant relationship with rainfall in northern than in northeastern Thailand. The highest value of the correlation coefficient is −0.60.
The relationship between positive IOD and rainfall showed the highest value of −0.75 for northern and −0.71 for northeastern Thailand. In contrast, MJO's significant correlation values were negligible. The significant correlations between El Niño, positive IOD and rainfall, maximum temperature, vertical mixing ratios, and vertical temperature demonstrated the probability of severe drought exacerbation in a dry spell and dry season. The moderately significant correlations between El Niño and vertical mixing ratios with the highest values of −0.63 in March indicate that they coincide with El Niño and rainfall showing significant correlations in March and April in northern Thailand.
The significant correlations between La Niña, negative IOD, and regional weather variabilities were found less than in El Niño, positive IOD, and found more in northern than in northeastern Thailand. The results suggest a potential for improving teleconnection-induced meteorological event forecasts in northern and northeastern Thailand. Since teleconnections may exacerbate extreme events such as droughts, additional policy planning for water resource management, extreme events risk management, and agriculture should be considered to reduce adverse consequences on the environment, economics, and livelihoods.
Future research could further examine the relationship between teleconnections and climate variables using additional climate models or ensemble data, especially vertical water vapor data for more reliable patterns. Further advanced statistical methods should be applied to find more associations and impacts between teleconnections and MJO vs regional weather such that robust forecast modeling with ensemble teleconnections can be established.
ACKNOWLEDGEMENTS
The Ministry of Agricultural and Cooperatives, Thailand, supports the first author through a graduate scholarship. Furthermore, the first author would like to thank Mr Chanti Detyothin, Expert on Applied Atmospheric Science Research and Development, Department of Royal Rainmaking and Agricultural Aviation, Ministry of Agricultural and Cooperatives, Thailand, who provided valuable insights and expertise that greatly assisted the research. The authors would like to thank Thai Meteorological Department and Hydrology Division, Office of Water Management and Hydrology, Royal Irrigation Department Thailand for the valuable climate data.
DATA AVAILABILITY STATEMENT
Data cannot be made publicly available; readers should contact the corresponding author for details.
CONFLICT OF INTEREST
The authors declare there is no conflict.