Impact of El Niño, Indian Ocean dipole, and Madden–Julian oscillation on land surface temperature in Kuching City Sarawak, during the periods of 1997/1998 and 2015/2016: a pilot study

The years 1997/1998 and 2015/2016 marked the most intense El Niño events ever recorded in human history. These occurrences of the El Niño Southern Oscillation (ENSO) phenomenon are frequently associated with devastating droughts and floods (Kovats et al. 2003). ENSO consists of three distinct phases: El Niño, characterized by warm tropical Asia-Pacific sea surface temperatures; La Niña, characterized by cold tropical sea surface temperatures; and Neutral, which is neither El Niño nor La Niña. Each event possesses unique characteristics (Trenberth & Stepaniak 2001). Numerous studies have investigated the correlation between ENSO and temperature, including the works of Tan et al. (2020), Tavakol et al. (2020), and Eboy & Kemarau (2023). In Malaysia, the impact of the ENSO on climate variables is well documented and significant, contributing to notable climate variability (Tangang et al. 2012). Despite this, research remains sparse on how the Madden–Julian oscillation (MJO) and the Indian Ocean dipole (IOD), alongside ENSO, collaboratively influence local temperature fluctuations and broader climate patterns. This gap underscores the necessity of focused investigations into the combined effects of MJO, IOD, and ENSO on local land surface temperatures (LSTs), which could provide crucial insights for urban climate management and planning.

The decision to focus on a localized region for this pilot study is twofold. First, the unique geographical and climatic conditions of the Borneo region make it particularly susceptible to the effects of the IOD, MJO, and ENSO. This susceptibility provides a compelling case for studying the complex interplay and impacts of these climatic phenomena on local LST. Second, by conducting a focused study in a specific region, the aim is to establish a detailed and robust methodology that can be replicated or adapted for broader geographical analyses in future research. This approach not only fills a significant research gap by providing in-depth local data but also tests the effectiveness of integrating various climatic data sources on a smaller scale before their broader application.

In a comprehensive literature review, Islam et al. (2018) and Nurdiati et al. (2022) explored the effects of ENSO, IOD, and MJO on precipitation and temperature, considering the relevance of biomass burning. Lenka et al. (2022) focused on the impact of these phenomena on the Indian Summer Monsoon, and Huang et al. (2022) investigated extreme IOD events associated with ENSO and MJO at a regional scale. The complex relationship between El Niño, IOD, and MJO was highlighted due to their diverse spatial patterns (Huang et al. 2022). However, there is limited understanding of the influence of MJO, IOD, and ENSO on human comfort in urban areas, despite their significant contributions to climatological and meteorological anomalies such as heatwaves affecting urban populations (Garreaud et al. 2019; Tan et al. 2020; Eboy & Kemarau 2023). While some comprehensive studies have covered the entirety of Malaysia (Ibadullah et al. 2019; Tan et al. 2020), the relationship between the intensity of MJO, IOD, and El Niño events and local weather and climate remains limited, particularly in the context of satellite remote sensing (RS).

Most RS studies have focused on developed countries, particularly in high-latitude regions, analyzing sea surface temperature (SST), analyzing the impact of land use change on LST and rainfall (Tangang et al. 2012; Houk et al. 2017; Sharma et al. 2020, 2023). As each El Niño event exhibits distinct intensity, formation processes, and duration, there is considerable interest in comparing the differences between the MJO, IOD, and El Niño events of 1997/1998 and 2015/2016 (Houk et al. 2017). Further research in this area can provide valuable insights into the intricate interactions between climate phenomena and their impact on local weather and climate conditions. Previous research has highlighted that the impacts of El Niño events can vary based on the geographical location of a region. Malaysia's location within the Maritime Continent (MC) makes it particularly susceptible to the ocean–atmospheric interactions influenced by ENSO, IOD, and MJO, as emphasized by Eboy & Kemarau (2023). These phenomena play crucial roles in shaping the climate patterns of the MC region. The significance of the MC lies in its position at the center of the tropical warm pool in Southeast Asia, making it a focal point of interest for researchers. The intricate interplay between the ocean and the atmosphere in this area significantly impacts monsoon activity and weather patterns. The eastern regions of Peninsular Malaysia, along with Sabah and Sarawak, are significantly impacted by the ENSO, which has been shown to increase the incidence of dryer-than-normal conditions by 20–30% (Mohmud 2018). These areas, prone to variable climate conditions influenced by ENSO, also experience interactions with the IOD and the MJO, though the collective impact of these systems on local temperatures and climate anomalies remains underexplored.

Addressing this research gap, this study focuses on these specific regions, which are critically affected by these climatic oscillations, to better understand their combined effects on LST. This knowledge is crucial for effective climate adaptation strategies, particularly in urban areas where temperature fluctuations have direct implications on human comfort and urban infrastructure. Researchers are particularly keen on understanding how ENSO, IOD, and MJO influence the climate of the MC, given their significant consequences for weather systems in Malaysia and the broader region. By exploring these influences, scientists aim to gain valuable insights into the region's climate variability and enhance the ability to predict and adapt to weather changes. In the context of the tropically situated study area, factors such as the monsoon, IOD, and MJO have a strong influence. Tan et al. (2020) and Kemarau & Eboy (2021a, 2021b, 2021c) have utilized temperature data from the Malaysia Meteorological Department (MMD) to demonstrate the impact of El Niño on temperature. However, these studies have not provided spatial information regarding the localized patterns of LST. In other words, while previous studies have acknowledged the impact of El Niño on temperature using MMD data, they have not specifically addressed the spatial distribution of LST at a local scale. This spatial information is crucial for understanding the localized effects of El Niño on LST patterns in the area of interest. The objective of this study is to examine the differences in oceanic and atmospheric conditions during extreme ENSO events and explore the synergy between these climate phenomena to provide a comprehensive picture of their climatic influence.

The dataset detailed in Table 1 comprises Landsat imagery, meticulously selected for clarity and absence of cloud coverage, essential for the accurate retrieval of LST and land cover data. The selection specifically targets years corresponding to significant El Niño events 1997/1998 and 2015/2016 due to their pronounced impact on regional climate conditions. This approach ensures the reliability and relevance of the RS data utilized in the analysis.

The Oceanic Niño Index (ONI), a reliable indicator for identifying El Niño and La Niña events, is widely recognized and accessible from the National Oceanic and Atmospheric Administration (NOAA) website at NOAA's ONI page (Huang et al. 2016). It tracks the development and intensity of El Niño or La Niña in the Pacific Ocean by calculating SST anomalies observed in the Niño region 3.4 (5°N–5°S, 120°W–170°W). El Niño is defined when the average SST anomaly over 3 months is at or above +0.5 °C, while La Niña is defined when the average SST anomaly is at or below −0.5 °C (National Oceanic & Atmospheric Administration Forecast Center 2019). ENSO events are categorized into five classes based on their intensity: weak (anomalies of 0.5–0.9 °C), moderate (1.0–1.4 °C), strong (1.5–1.9 °C), and very strong (≥2.0 °C) for El Niño events, with opposite classifications for La Niña events. Data accessed from the National Oceanic and Atmospheric Administration Climate Prediction Center.

The IOD is a significant climatic pattern that influences weather conditions across the Indian Ocean region. Relevant datasets for the IOD can be accessed at NASA's Sea Level Portal, available at https://sealevel.jpl.nasa.gov/overlay-iod/, which provides comprehensive data essential for the analysis. In its positive phase, warm waters are pushed toward the western part of the Indian Ocean, while cold deep waters rise to the surface in the eastern Indian Ocean. The pattern is reversed during the negative phase. The intensity of the IOD is measured by the anomalous SST gradient between the western equatorial Indian Ocean (50°E–70°E and 10°S–10°N) and the southeastern equatorial Indian Ocean (90°E–110°E and 10°S–0°N). This gradient is known as the dipole mode index (DMI). A positive DMI indicates a positive IOD, while a negative DMI indicates a negative IOD. The DMI data can be obtained from the National Oceanic and Atmospheric Administration Working Group on Surface Pressure.

The pre-processing stage involves several crucial steps, including geometric correction, atmospheric correction, and radiometric correction, to ensure their usefulness for temperature research (Sharma et al. 2023). Subsequently, the surface temperature of the thermal band is extracted for the specific datasets from 1997/1998 and 2015/2016. The process for inverting surface temperature from Landsat 5 and Landsat 8 imagery, available for download at https://earthexplorer.usgs.gov/, adheres to the methodologies established by Kemarau & Eboy (2019). This procedure ensures accurate temperature retrieval, which is essential for the analysis.

Table 2 presents waveband information for the Landsat 8 Operational Land Imager (OLI) thermal infrared sensor (TIRS) satellite. The Landsat Satellite 8 follows a sun-synchronous orbit at an altitude of 705 km near the Earth's poles. It is equipped with the Instrument OLI consisting of nine spectral bands and two thermal infrared sensors, as detailed in the table.

Moreover, this study integrates data from the Landsat 8 OLI TIRS, Landsat 7 ETM + , and Landsat 4/5 TM RS systems. For detailed information regarding the paths associated with Landsat 5 TM, please refer to Table 3.

According to Equation (1) provided, the variables are defined as follows: represents the pixel value as radiance in W/(m² sr μm), LMAX_λ is the spectral radiance scale factor to Q_calmax in W/(m² sr μm), and LMIN_λ is the spectral radiance scale factor to Q_calmin in W/(m² sr μm). Q_calmax corresponds to the maximum calibrated pixel value (equivalent to LMAX) in DN, which is 255, while Q_calmin corresponds to the minimum calibrated pixel value (equivalent to LMIN) in DN (Table 4).

Table 4

LMAX and LMIN information for Landsat 5 TM

Information .	Value .
	15.303
	1.238

Information .	Value .
	15.303
	1.238

View Large

In this study, the radiance spectrum is measured in units of W/(m² sr μm), while ML represents the recalibrated gain for a specific band, also in units of W/(m² sr μm). ML corresponds to the path radiance, measured in units of W/(m² sr μm). The ENVI 5.3 software is employed, which provides thermal functions specifically designed for atmospheric correction purposes. Before conducting atmospheric correction, it is necessary to convert the thermal images into radiance format.

The term ‘radiance’ refers to the observed radiance at the aperture, while τ represents the atmospheric transmission. B(T) corresponds to the Planck function, and Lu represents the upwelling radiance due to the atmosphere. The ISAC algorithm evaluates the surface radiance emitted based on the brightness temperature of each pixel. The scatter plot of radiance for each band follows the Planck formula. A linear regression line is fitted to the upper portion of the scatter plot, which represents pixels approaching an emissivity of 1.0. This process estimates the transmission (slope) and upwelling (intercept) factors. These factors, referred to as ‘transmission’ and ‘upwelling’, approximate the surface radiance leaving the Earth for each pixel within the original cube. By plotting the transmission and upwelling (τ, Lu) and generating the cube, the ENVI 5.3 algorithm incorporates this information to perform thermal atmospheric correction using ISAC.

Surface temperature is typically measured in Kelvin (K). In the context of the mentioned equation, T_B represents the effective sensor brightness temperature in K. The variable λ denotes the wavelength of the emitted light, expressed in meters. For the Landsat 5 TM thermal bands, the effective wavelengths are 11.457, 11.269, and 10.904 μm, based on Equation (6) with α = 1.438 × 10⁻² mK. The symbol ε represents the surface emissivity.

The thermal surface radiance is captured by the TIRS band 10 on Landsat 8 OLI. Pre-processing is conducted to eliminate atmospheric effects from the thermal data, and the surface temperature is derived using the algorithm proposed by He et al. (2019).

In Equation (7), L_λ represents the spectral radiance in units of W/(m sr μm)². ML corresponds to the rescaled gain specific to a particular band, measured in units of W/(m sr μm)². A_L accounts for the band-specific rescaling bias, also measured in units of W/(m sr μm)². It is important to note that the rescaling bias is applied to a specific band.

In Equation (8), K_1 and K_2 are predefined constants set to 774.89 (in W/(m sr μm)²) and 1,321.08 (in W), respectively. The surface temperature is determined by correcting the emitted radiance of the ground (B) in Kelvin (K) using the mono-window algorithm proposed by Qin et al. (2001).

In Equation (10), ε represents the land emissivity, εv represents the emissivity of pure vegetation (set to 0.985), εg represents the emissivity of bare soil (set to 0.960), and (A parameter representing the adjustment to an average value of 0.01, used to refine emissivity calculations based on specific environmental conditions) is a parameter adjusted to an average of 0.01.

In Equation (11), T_B represents the sensor brightness temperature in Kelvin (K), λ is wavelength of the radiation (specific to the sensor's operational band), C2 is the second radiation constant in Planck's law equation, ε is the overall emissivity used in the conversion from brightness temperature to actual surface temperature, and ρ = hc/σ, where σ is the Boltzmann constant (1.38 × 10⁻²³ J/K), h is the Planck constant (6.626 × 10⁻³⁴J/s), c is the speed of light (2.998 × 10⁸m/s), and ε is the emissivity.

After generating the LST maps, a correlation analysis were performed between the LST values and the mean temperature recorded between 1,400 and 1,500 h on the dates coinciding with satellite captures in the study area. Meteorological data from the MMD were used for this analysis. It is important to note that the pixel coordinates in the LST maps correspond to the locations of the meteorological stations to validate the LST from the Landsat satellite.

Next, the spatial patterns of the LST maps were compared between the El Niño years of 1997/1998 and 2015/2016. This comparison aimed to identify any temperature distribution differences during these two El Niño events.

Moreover, this study performed statistical comparisons on the mean, maximum, and minimum LST values extracted from the LST maps of the years 1997/1998 and 2015/2016. These statistical analyses yielded significant insights into the changes in temperature characteristics associated with the two El Niño events. To facilitate the interpretation of these comparisons, graphical representations were employed.

Data of MMD (temperature C), MJO, IOD, and ONI were compiled and synched based on date, from February 1993 until December 2020. The compiled data were imported into R programming via RStudio and transformed into a time-series data format. MMD was treated as the dependent variable, while MJO, IOD, and ONI were treated as co-variate. The time-series data were decomposed using ggsdc (ggseas V0.5.4) to observe the trend, seasonality variation, and noise frequency in each series. A normality test (Shapiro–Wilk test) was conducted on the MMD to examine the distribution pattern of the dependent variable (MMD). A generalized linear mixed model (GLMM) using Template Model Builder was constructed to test if the interactions of MJO, IOD, and ONI were significant in driving the fluctuations of MMD. MMD was set as a Gaussian distribution for GLMMTMB.

The research team employed the ISO Cluster classification method to develop land cover maps for the years 1998 and 2016, chosen due to the availability of the clearest satellite images with minimal cloud cover. This clustering algorithm grouped pixels with similar spectral characteristics into various land cover types, aiding in the identification and characterization of different land cover within the study area. The precision of these maps was assessed using a time-series analysis from Google Earth, following the methodology outlined by Kemarau & Eboy (2020). This analysis involved comparing the land cover maps with high-resolution satellite imagery and other reliable sources to ensure the accuracy and reliability of the classifications. The primary goal of the study was to categorize the data into four main surface cover types: urban areas, vegetation, wetlands, and water bodies. The validation process was crucial for assessing the reliability of RS data in identifying hotspot prone areas, particularly during dry and heatwave events. This endeavor required close collaboration with relevant departments to ensure effective assistance and intervention during such critical weather occurrences. The findings obtained through this comprehensive approach were expected to offer valuable insights into the intricate interactions between climate phenomena and LST, ultimately contributing to an enhanced comprehension of climate variability in the study area.

The correlation analysis yielded a correlation coefficient of 0.97, indicating a strong relationship between the LST values and the ground-based meteorological data. This result was consistent with the findings of Eboy & Kemarau (2023), who also reported a similar high correlation between satellite-derived LST and ground-based measurements. Notably, the pixel coordinates in the LST maps align with the location of meteorological stations, enabling the validation of LST data obtained from the Landsat satellite.

The patterns and influences of the ONI, MJO, and IOD on the mean monthly temperature (MMT) in the study area are examined in this study (Figure 2). The ONI values serve as indicators of the intensity of El Niño events during the specified years. Positive ONI values (above 0.5) indicate the presence of El Niño, which is associated with warmer sea surface temperatures. Conversely, negative ONI values (below −0.5) indicate the presence of La Niña, associated with cooler sea surface temperatures. The table shows that during the El Niño years (January 1997–December 1998 and January 2015–December 2016), the ONI values range from −1.6 to 2.6. Notably, there is a trend of increasing ONI values from negative to positive during the El Niño periods, signifying a transition from La Niña to El Niño conditions. This transition is accompanied by a gradual increase in MMT in the study area.

The MJO values in the table represent the intensity of the MJO, an atmospheric phenomenon that impacts tropical weather patterns. The MJO index indicates the presence of enhanced or suppressed rainfall in the region. Throughout the specified years, the MJO values vary. Some months exhibit higher MJO values (>1), indicating a stronger influence on weather patterns, while other months have lower values (<1), suggesting a weaker impact. The influence of MJO on MMD is not straightforward and can vary depending on the specific phase and intensity of the MJO. However, it is generally expected that during periods of enhanced rainfall (high MJO values), there may be a cooling effect on MMD due to increased cloud cover and precipitation.

The IOD index also shows significant variability but with less frequency and amplitude compared with the MJO. Similar to the MJO, the IOD does not display a clear long-term trend, suggesting that the dipole effect remains consistent without significant changes in intensity over the years. The ONI, which measures the ENSO phenomena, shows periodic spikes that correlate with known El Niño events, particularly noticeable in years such as 1998, 2010, and 2016. These spikes represent the warm phases of ENSO, known as El Niño, which typically lead to higher global temperatures. There are dips corresponding to the La Niña events, which are cooler phases, though they are less pronounced in this graph. The temperature, represented by MMD (mean monthly temperature), generally trends upward over the observed period, with notable increases aligning with El Niño events (e.g., the peaks around 1998 and 2016). This suggests a correlation between high ONI values and increased regional temperatures. The overall upward trend in temperature across the years could be indicative of longer-term warming trends, possibly influenced by global climate change in addition to ENSO cycles.

Table 5 from the GLMMTMB analysis provides detailed regression coefficients along with their statistical significance for various climate indices and their interactions, focusing on their effects on MMT. The intercept of 26.69622, with a z-value of 444.5 and a highly significant p-value (<2e − 16), indicates the base level of MMD when all other indices are at zero, showcasing the model's robust baseline for temperature predictions.

The coefficient for the IOD stands at 0.38867, demonstrating a positive and statistically significant relationship (p-value = 0.0455) with MMT, suggesting that higher IOD values correlate with increased temperatures. In contrast, the ONI has a smaller, non-significant positive effect (coefficient = 0.057, p-value = 0.4978), indicating that ONI's influence on temperature, while present, does not consistently impact MMT. Similarly, the MJO shows a negative coefficient (−0.02927) with a non-significant p-value (0.5552), highlighting its minimal and inconsistent effect on temperature changes.

Regarding interaction terms, both IOD:ONI and IOD:MJO interactions display negative coefficients (−0.23575 and −0.29962, respectively) but lack statistical significance, hinting at possible moderating effects where the presence of one index might reduce the impact of another on temperatures. The ONI:MJO interaction, with a positive coefficient of 0.1325, suggests a potential synergistic effect that nearly reaches statistical significance (p-value = 0.0519), possibly indicating the combined influences of these indices under specific conditions. Lastly, the three-way interaction (IOD:ONI:MJO) shows a positive effect (0.11724), though it is not statistically significant (p-value = 0.5508), suggesting that the combined influence of these indices on temperature is complex and requires further investigation. Overall, this regression analysis underlines the nuanced relationships between climate indices and regional temperatures, with the IOD showing a clear impact, while the effects of ONI and MJO are more variable. These findings are critical for refining predictive models and enhancing the understanding of how these climate phenomena influence regional climate patterns. However, during extreme events, particularly El Niño 1997/1998 and 2015/2016, when the ONI values spiked significantly higher compared with IOD and MJO, these indices substantially influenced MMT (Kemarau & Eboy 2019) and LST (Eboy & Kemarau 2023; Kemarau & Eboy 2023).

During positive IOD and El Niño periods, both the Asian continent and Western Australia experience drought-like conditions. Previous studies have highlighted that in drought years, the SST in the East Indian Ocean is significantly lower (Streten 1983; Lenka et al. 2022). It is well established that ENSO phases strongly influence Australian rainfall, with El Niño events leading to below-average precipitation (Ropelewski & Halpert 1987; Lenka et al. 2022). However, with the discovery of IOD, the understanding of annual Australian rainfall has evolved slightly. Observational studies have shown that despite being in an El Niño period, widespread and substantial anomalies occur due to overlapping positive IOD years (Nicholls 1989).

Several researchers have investigated the relationship between local MJO in the Pacific and ENSO SST indices, revealing strong correlations (Kessler & Kleeman 2000). However, in the global context, no such relationship has been reported, which is consistent with the findings of this study, as supported by Slingo et al. (1999) and Hendon et al. (1999). Due to the MJO, oceanic Kelvin waves and warm surface water travel eastward, enabling the MJO to extend further into the central Pacific. This phenomenon occurs concurrently with El Niño phases (Hendon et al. 1999; Lenka et al. 2022), but it does not significantly impact LST, as found in this study. Some studies suggest that intra-seasonal variability in MJO circulation induces tropospheric precipitation, which may directly influence deep convective conditions (Tompkins 2001) or indirectly affect moist stability (Myers & Waliser 2003). Similarly, SST also influences MJO, as mean SST determines MJO propagation speed and period by maintaining a low moisture level (Davey 1989). To analyze the influence of IOD phases on different regions, researchers have conducted studies using multi-source observed data and model simulations (Kayano et al. 2021; Sena & Magnusdottir 2021) to assess the combined processes.

Table 6 presents the spatial changes in LST between the El Niño events of 1997/1998 and 2015/2016, as shown in Figure 7. The results indicate notable differences between the two events. In terms of mean temperature, El Niño 2015/2016 recorded a higher value compared with El Niño 1997/1998. Similarly, the maximum temperature during the 2015/2016 El Niño event was also higher than that of the 1997/1998 El Niño event. However, the minimum temperature values varied depending on the ONI value. For instance, during El Niño 1997/1998, the minimum temperature occurred at an ONI value of 1.9, while for El Niño 2015/2016, it occurred at an ONI value of 1.

These findings align closely with the study conducted by Houk et al. (2020), which observed a higher temperature value of 3.5 °C during the El Niño 2015/2016 event due to warming in the central Pacific Ocean, surpassing the 1997/1998 event. The study also revealed that the degree of El Niño's effect depends on the different phases of El Niño formation, particularly the rate of increase and decrease of ONI. In addition, Thirumalai et al. (2016) noted that April 2016 recorded the highest LST in history, coinciding with the peak of the 1997/1998 El Niño event. They further highlighted the contribution of global warming to rising temperatures. These findings support the results of the current study.

Despite the significant influence of El Niño, it is important to consider the impact of urbanization activities on the increase in LST (Eboy & Kemarau 2023). This explains the higher mean and maximum LST values observed during the El Niño 2015/2016 event. Furthermore, Thirumalai et al. (2016) found that April 2016 set a record for the highest temperature in the history of El Niño. Table 7 illustrates LST data for different years under varying conditions of three key climate indices: ONI, IOD, and MJO. The years selected – 2016, 1998, 2018, and 2011 – represent different climatic anomalies as characterized by the values of ONI, IOD, and MJO.

The year 2016, with ONI at 2.4, represents a strong El Niño phase, which typically results in higher temperatures globally (Figure 2). The maximum LSTs recorded in urban and industrial areas (35.81 and 35.9 °C, respectively) align with expected increases due to El Niño's influence on heat accumulation in built environments. Vegetation and wetlands show slightly lower temperatures, indicating their cooling effect even under El Niño conditions. This year also exhibits an El Niño condition, though with a positive IOD value (0.8). The combination of a positive IOD with El Niño can enhance warming effects in certain regions, particularly Southeast Asia and Australia (Weller & Cai 2013). The LSTs are generally high but slightly lower than in 2016, potentially due to the mitigating effects of a less intense El Niño and the regional impacts of the IOD.

With negative values for both ONI and IOD, 2018 reflects La Niña conditions, typically associated with cooler, wetter weather in the equatorial Pacific region. This is reflected in the generally lower maximum, mean, and minimum LSTs across all types of land cover compared with El Niño years. The cooling effect is particularly evident in water bodies and wetlands, which register the lowest temperatures, emphasizing the significant influence of La Niña on water-rich environments. Similar to 2018, 2011 was a year of La Niña conditions but with very weak MJO activity. The generally low LSTs across all categories, especially in water bodies and wetlands, corroborate the expected cooling effects of La Niña. The weak MJO might have contributed to less variability in temperature, resulting in consistently lower temperatures across the board.

The data show a clear pattern where El Niño years are associated with higher temperatures across most land covers, particularly in urban and industrial areas, which are more susceptible to heat due to the urban heat island effect. Conversely, La Niña years show a general cooling trend, especially noticeable in natural landscapes such as water bodies and wetlands. These observations are consistent with the known effects of these climate phenomena on global and regional climate patterns, as documented in the literature, where El Niño events have been shown to increase temperatures by 0.5 − 1.5 °C according to data provided by weather stations in Kuching (Panisset et al. 2017; Moura et al. 2019; Kemarau & Eboy 2023). Eboy & Kemarau (2023) also note that rainfall plays a critical role in shaping local temperature patterns, reporting that El Niño contributes to drought and dry conditions due to reduced rainfall, factors that contribute to temperature increases in Kuching during these events.

Furthermore, Mahmud (2018) reported that El Niño decreases rainfall by 25–30% compared with neutral conditions. Surface cover types such as wetlands, water bodies, vegetation, and urban, and industrial areas exhibit higher temperatures during El Niño compared with La Niña. Consistent with previous scientific reports, industrial areas have the highest temperatures due to heat generation and significant greenhouse gas emissions (Li et al. 2011; Kemarau & Eboy 2023). Figure 8 details changes in surface temperature patterns based on the type of land cover.

According to Figure 8 and Table 7, water bodies such as the Sarawak River and vegetation have lower temperatures compared with urban areas. For instance, water bodies are around 26 °C, and vegetation varies from 27.5 to 29 °C, whereas urban areas exceed 29 °C. Temperatures depend on the volume of vegetation, its quantity, and the extent of area it covers. Urban areas experience the highest temperatures due to the presence of man-made structures that absorb and release heat, thus raising temperatures (Estoque & Murayama 2017; Kemarau & Eboy 2023). The warming of urban areas, consisting of man-made materials, has become a significant contributor to urban heat islands, especially when indoor air conditioning is used (Parker 2010). Kemarau & Eboy (2023) also report that emissions of carbon dioxide (CO₂) and carbon monoxide (CO) in industrial zones cause local temperature increases.

According to Figure 9, it is clear that the hotspot area during the El Niño event exceeds that during the neutral event. Specifically, during the El Niño events of 2015 and 2016, the hotspot areas covered 95.55 and 97.8 km², respectively. In contrast, during the neutral events in May 2014 and October 2017, the hotspot areas measured 66.8 and 73.3 km², respectively. This consistent pattern of concentrated heat in the region persisted during the 1998 El Niño event, with a hotspot area of 89.32 km² (Eboy & Kemarau 2023).

These findings provide evidence of the influence of ENSO on the alteration of hotspot patterns. The observed increase in a hotspot area, ranging from 34 to 36 km², during the El Niño event, corresponds to a rise in electricity consumption of 20–30% above normal levels. This increase in electricity usage is aimed at balancing temperatures in residential and office areas in Kuching City, Sarawak. These results align with previous studies conducted by Lin et al. (2018) and Drosdowsky & Wheeler (2014), which have highlighted the contribution of El Niño events to global temperature rise. This global temperature increase indirectly impacts local temperatures, particularly in the study location. Furthermore, supporting evidence for this phenomenon can be found in the works of Tan et al. (2020) and Moura et al. (2019). Yu et al. concluded that the influence of local climate does not affect the impact of dominant El Niño warming, especially during the mature stage.

This study has elucidated the significant impacts of the IOD, MJO, and ENSO on LST in the MC, specifically in Kuching, Sarawak, Malaysia, during the severe El Niño events of 1997/1998 and 2015/2016. The key findings indicate that the ONI plays a pivotal role in modulating temperature extremes and dry spells, exerting a pronounced influence during peak El Niño years. In addition, the IOD also exhibits a considerable, albeit less pronounced, impact on regional temperatures. These findings have substantial implications for climate research, highlighting the need for integrated climate monitoring systems in the region to better predict and mitigate the effects of such climate phenomena. For policymakers, this information underscores the importance of developing and implementing more effective climate resilience and adaptation strategies, especially in areas highly vulnerable to ENSO-driven anomalies.

Considering the complexities observed in the interactions between these climate systems, future research should focus on several areas. Enhanced spatial analysis is needed to provide more detailed investigations that could help delineate the localized impacts of these phenomena, potentially uncovering microclimatic variations critical for urban planning and agricultural practices. Furthermore, longitudinal studies involving extended time-series analyses spanning multiple climate cycles could offer deeper insights into the long-term trends and shifts in climate patterns due to global climate change. Lastly, applying interdisciplinary approaches that integrate ecological, socio-economic, and climatological data could enhance the understanding of the broader impacts of climate variability on ecosystems and human communities in the MC. This study's approach and findings contribute valuable knowledge to the field of tropical climatology and provide a solid foundation for further explorative studies aimed at refining the predictive capabilities and response strategies to climate variability.

We acknowledge NASA for providing Landsat data and the Malaysia Meteorological Department for providing meteorological data.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.