Spatiotemporal characteristics of hydro-meteorological droughts and their connections to large-scale atmospheric circulations in the Kelantan River Basin, Malaysia

Droughts have significantly impacted a wide variety of socioeconomic activities such as water supply, agriculture, aquaculture, industry, and hydropower generation (Munyasya et al. 2022). In Southeast Asia, droughts have frequently destroyed crops, caused forest fires and depleted water supplies. According to the International Disaster Database, 52 drought episodes have affected more than 77 million people in Southeast Asia since 1996 (EM-DAT 2018). Drought is defined as water deficiencies when compared to normal circumstances, where it can be classified as hydrological, agricultural, meteorological, and socioeconomic, depending on the impact. According to Leta (2018), droughts will become more frequent, severe, and persistent in many parts of the world. While Siderius et al. (2018) also stated that the frequency and duration of severe drought events are predicted to increase in the future. This is because, based on the United Nations, the human population will increase by 2 billion in the next 30 years, and the population growth leads to more resources extraction and consumption that accelerate climate change.

In Malaysia, precipitation characteristics are influenced by two monsoon seasons, the northeast monsoon (NEM) from November to March and the southwest monsoon (SWM) from May to September. While the inter-monsoon normally occurred in April and October. During these monsoon seasons, total annual precipitation between 2,000 and 4,000 mm with 150–200 wet days, falls over the nation (Suhaila & Aziz 2007). However, a reduction in total precipitation and the number of precipitation days during the SWM season in Peninsular Malaysia, which might potentially cause issues in the agricultural, industrial, water supply, public health and energy sectors (Suhaila et al. 2010). Other than that, Hasan et al. (2021) applied Streamflow Drought Index (SDI) to study the spatiotemporal pattern of drought in Peninsular Malaysia across 3, 6, 9, and 12 months. They found that 1997–1999, 2002, and 2016–2018 were severe drought periods in Peninsular Malaysia. Short-term droughts are commonly occurred in most areas, while more severe droughts can be found in the northeast and southeast regions. For instance, a severe drought in 2014 impacted paddy productivity in Kelantan located in the northeastern part of Peninsular Malaysia, resulting in approximately USD$22 million losses in agricultural industry (Tan et al. 2017). In addition, Tew et al. (2022) also reported rapidly rising temperatures and high variability rainfall patterns in Kelantan, which could eventually cause severe droughts. For example, some severe droughts were found in 1997, 1998, 2002, 2003, 2005, 2006, 2007, 2009, and 2010 (Yusof et al. 2013; Hashim et al. 2016).

Standardized Precipitation Index (SPI) and Standardized Streamflow Index (SSI) have been utilized to determine the severity, duration, and intensity of basin-scale drought using precipitation and streamflow data, respectively (Tan et al. 2019). SPI was created to define and track droughts by McKee et al. (1993). SPI is simple to apply since requires only monthly precipitation data and is able to compute various timeframes and widely used by other researchers. For example, Luhaim et al. (2021) utilized SPI-3 to study the characteristics of drought events in the Muda River Basin, where they found the basin was affected by severe droughts in 1991–1992, 1995, 1998, 2002–2003, 2005–2006, 2008, 2012–2013, and 2016. Other than that, they also found the SPI values decreased throughout the three time periods, which is indicative of an increasing trend of drought events across the basin. Climate change may also account for the increase in the amount of dry months at most of the rainfall stations in the Sarawak River basin during the last decade compared to the preceding 30–40 years (Bong & Richard 2020). Almost all the stations had at least 5 very dry periods, and most of the stations endured roughly 10 severely dry droughts. Eventually, information may assist enhance the system for early warning of drought and drought mitigation techniques. Salimi et al. (2021) have utilized SPI, SSI, and Standardized Precipitation Evapotranspiration Index (SPEI) to analyse the dry and wet periods across three basins in Iran. Moreover, Lee et al. (2017) due to climate change, droughts were more severe between the years 1995 and 2085 in Hwanghae Plain, North Korea. The SPIs revealed that future 1-month droughts would be more severe than future 6-month droughts, whereas the SPEI showed the reverse. The findings demonstrate that the association between the meteorological and hydrological dryness is substantial at 99% in all three basins. However, an in-depth analysis of correlation between meteorological and hydrological droughts as well as their relationship with large-scale atmospheric circulation is relatively limited.

This study aims to assess the spatiotemporal characteristics of droughts in the KRB from 1985 to 2020. The specific objectives of this study include: (1) to quantify the hydro-meteorological drought characteristic, i.e., duration, peak, severity, and intensity; (2) to evaluate the response rate of SSI to SPI; and (3) to analyse the impact of large-scale atmospheric circulations on the drought formation in the KRB. The findings are important for local stakeholders to formulate regional water resources management and disaster reduction plans. For a stronger preventive measures and readiness, planning authorities must identify temporal and geographical patterns of existing droughts for forecasting future droughts. Lastly, the findings could contribute to improve the current literature on the tropical droughts and how large-scale atmospheric circulation, i.e., El Niño Southern Oscillation (ENSO) affect the drought formation.

Monthly precipitation data from 1985 to 2020 at 21 meteorological stations installed in the basin were collected from the Malaysian Meteorological Department (MMD). Monthly streamflow data for the same period at the Jambatan Guillemard was collected from the Department of Irrigation and Drainage (DID) Malaysia. Basic information and distribution of the meteorological and hydrological stations is shown in Figure 1.

World Meteorological Organization (WMO) suggests the use of SPI for characterizing droughts. SPI values below −1 indicate severe precipitation shortages, whereas values above 1 indicate severe precipitation excesses. Table 1 lists the several types of wetness and dryness based on the SPI values. For example, a severely dry condition is defined when the SPI value is less than −2.00. In Malaysia, relatively dry periods typically last for not longer than 3 months (Tan et al. 2019), so SPI three-month time scale (SPI-3) was selected for the drought assessment in this study. The key advantage of SPI established by McKee et al. (1993) is that it can study drought at multiple time scales for different forms of drought. In this study, SPI-3 was calculated using the SPI tool created by the National Drought Mitigation Centre (NDMC). SPI-3 is suitable to analyse the moisture conditions for short and medium-term time scales and, therefore, accessible to offer a seasonal estimate of precipitation (Haroon et al. 2016).

While SPI uses monthly precipitation data, SSI uses data on monthly streamflow. SSI is computed based on the cumulative flow values estimated separately for each month, which is based on the terminology of SPI developed by McKee et al. (1993). SSI has been used in basin-scale drought study in Malaysia, i.e., Luhaim et al. (2021) to detect hydrological drought in the Muda River Basin, where most of the severe droughts endure at least 5 months. Types of droughts based on the SSI can be referred to in Table 1 as well.

ENSO is a major climatic phenomenon that influences severe weather conditions all over the globe. ENSO occurs every 5–7 years and is characterized by the warming of the seas caused by winds, resulting in the spread of heat over the planet. In this study, the National Oceanic and Atmospheric Administration (NOAA) model Nino 3.4 was employed to reflect ENSO fluctuation. It is a five-phase-averaged measure that identifies El Niño or La Niña occurrences. The index can freely be accessed from the NOAA website at https://www.cpc.ncep.noaa.gov/data/indices (accessed 20 August 2022).

The MJO is the dominant mode of intra-seasonal fluctuation in tropical regions. Generally, it flows to the east at around 4–8 ms1, entering the tropics in 30–90 days. Consequently, the MJO influences the intra-seasonal variability of monsoon precipitation. Past and present MJO activities can be quantified based on the outgoing longwave radiation (OLR) MJO Index (OMI), which is accessible at https://psl.noaa.gov/mjo/mjoindex/ (accessed 20 August 2022).

IOD is also known as the Indian Nino, occurs due to the difference in sea surface temperature between the eastern Indian Ocean and the Arabian Sea (Yue et al. 2021). Similar to ENSO, positive IOD values are associated with El Niño, while negative IOD values are associated with La Niña. The Dipole Mode Index (DMI), which can be retrieved from https://psl.noaa.gov/gcos_wgsp/Timeseries/DMI/ (accessed 20 August 2022), is generally used to represent the IOD.

D is the difference between the object's two rankings (drought index), and n indicates the sum of measurements.

Extreme dry events occurred in Gua Musang, Station Dabong, Kampung Laloh, and Ulu Sekor, at least 4 months a year. Besides, at least three extreme dry events were recorded in all stations from 1985 until 2020. Meanwhile, Balai Polis Bertam, Kampung Aring, and Brook experienced the most severe droughts, where the SPI values were less than −1.50, especially in 1987, 1989, and 2020. For example, Kampung Aring experienced severe dry spell from March to December 1989 and 2020. By the time, throughout the period, in 1986, 1987, 1989, 2008, and 2020, KRB had hit the highest number of occurrences of severe dry events. However, most of the stations experienced moderate drought for other years, for example, Kuala Krai in 1986 (March until June), 1995 (February, May, June, and December), and 2002 (March, June, October, and November). As a result, the highest number of dry events mainly occurred in 1986 and 2020.

Table 2 shows that between 1985 and 2020, the KRB experienced 13 dry spells based on SSI. More than 70% of dry spells last 6 months or longer. The longest drought (18 months) was seen between January 1997 and July 1998, followed by a 13-month period between November 2013 and December 2014 and from March 2005 until February 2006 (11 months). Even though the DP was the longest during events, the severity was only −11.36 in 1997/1998, while the highest severity was −13.87 in 2013/2014. There were only four extremely dry events, each with a peak of −2 or even less. Meanwhile, from October 1992 until November 1992, Jambatan Guillemard only had a lesser dry duration, with only one-month duration. Extremely dry periods, as defined by the SSI-3 with intensity level more than −1, were recorded in 1992, 1995–1996, 2004, 2005–2006, 2010–2011, 2012–2013, and 2013–2014. The most intense dry period was recorded from May 2010 to March 2011, with the intensity value of −1.28. The water level had decreased by over 2 m during the dry season in Kelantan, and to irrigate farms affected by drought, the Kelantan Drainage and Irrigation Department has released water from the Bukit Kwong dam (New Straits Time 2019). In fact, dam can effectively capture precipitation for storage and usage in the dry season (Chen et al. 2022; Wang et al. 2022; Zhu et al. 2022).

Other than that, Balai Polis Bertam and Kampung Aring had taken a place in the leading the occurrences of severely dry months, respectively. In contrast, moderately dry months have happened in most of the stations. Still, Kuala Krai had the highest number of moderately dry months compared to the extremely dry months. Besides that, Pej Jajahan Machang experienced a major variance among these three types of droughts. For example, the station had a higher number in moderately dry months, while only experienced one extreme dry month and four severely dry months. On the other hand, Balai Polis Bertam also faced a major difference between moderate, severe, and extreme drought months, e.g., in severe drought, the station experienced the more severely dry months than extremely and moderately dry months.

For further investigation, Spearman correlation analysis was utilized to measure the association between the annual series of monsoon indexes. The correlation values of the relationship between drought indices (SPI-3 and SSI-3) and large-scale atmospheric circulation (ENSO, MJO, and IOD) from 1985 to 2019 are listed in Table 3. In general, ENSO had a bigger impact on the meteorological and hydrological dryness in the KRB than IOD and MJO. ENSO showed a significant connection with SPI-3 in March (0.62), April (0.67), May (0.34), and June (0.36) at a 95% confidence level, showing ENSO encountered a favourable effect in the emergence of dryness in KRB during the dry season. Similarly, the response rate results show that DJF and MAM seasonal have the highest relationship. In fact, Kelantan receives low precipitation amount from February to April (Tan et al. 2017), where the impact on water supply could take part in the following months. Hence, the agricultural sector in this region was affected badly during the strong big El Niño years such as 1997–1998 and 2015–2016.

During the SWM phase, the eastern regions might be the driest and vice versa for the northwestern area (Suhaila et al. 2010). This is because the influence of the SWM is less than that of the NEM since Sumatra Island conceals by Peninsular Malaysia. Therefore, ENSO and SSI-3 demonstrated a significant correlation in February (0.32) and March (0.39), April (–0.55), and May (–0.43), which showed related to the dryness of that month. ENSO occurrences usually last between 9 and 12 months, whereas droughts in Malaysia often range between 2 and 6 months (Luhaim et al. 2021). Shaaban et al. (2011) stated that for the future intervals 2025–2034 and 2041–2050, the mean and maximum monthly streamflow at the same station will increase by a stated amount, but mostly during the months with high precipitation. By contrast, during 2015–2044, the monthly streamflow is anticipated to decline from June to October by a range of 1.1–9.1%, showing a drier condition might happen in the future.

At a confidence level of 95%, the MJO was linked with SPI-3 only in August (0.35), indicating that the phenomenon influences the meteorological drought creation during the middle phase of the SWM. With the correlation values of 0.46, 0.44, and 0.40, respectively, the IOD was connected to SPI-3 in March and SSI-3 in March and April at a 95% confidence level. This shows the dryness of the KRB in March influenced by both ENSO and IOD, which is consistent with the findings reported by Luhaim et al. (2021) in the Muda River Basin that located besides the KRB. Meanwhile, Tan et al. (2017) found that during 1985–2014, both consecutive wet days and consecutive dry days have decreased by 0.15 and 0.96 days/decade, respectively. In addition, the maximum 1-day and 5-day precipitation amount were also decreased from May to July and September to November. Similarly, they concluded that ENSO plays a crucial role in the climate system of Malaysia, and could bring prolonged drought events in the KRB.

Over the previous several decades, the KRB has experienced numerous severe dry events. For instance, the Kota Bharu and Kuala Krai stations experienced severe droughts most often. However, as Kelantan is identified as the Peninsular state with the highest flood risk, so lesser attention is given to the impact of dry spells in this region. Moreover, in this study, a similar result within Fung et al. (2020) experiment reported that in the East Coast and Northern areas, where precipitation and temperature have risen over the years. This clearly shows that KRB is still in a mild drought, and as a result, in SPI also, most of the stations are only in mild dry and more sensitive to brief periods of insufficient precipitation (dry spells). This is because Peninsular Malaysia has a tropical climate zone and is situated close to the equator, which may be the reason of the short-term dry events. The high temperatures and excessive precipitation are also a contributing factor. According to Adnan & Atkinson (2011), there could be seasonal fluctuation in the intensity of monsoon season rainfall in the Kelantan catchment. This study has proven that the seasonal fluctuation, particularly from February to May, could affect the water supply in the basin.

The agricultural sector in the KRB, particularly paddy, affected the most during the El Niño years due to insufficient water supply for irrigation. This has been further proven by the hydrological drought conditions as measured by SSI-3 using the observed streamflow data. These extreme weathers have significantly impacted the paddy farmers' income and livelihoods (Firdaus et al. 2020), affecting the national food security as rice is the staple food of Asia. To tackle the issue, more water treatment plants can be constructed in the regions that affected by less intense drought, i.e., along the Lebir River in the southeastern part and in the middle part of the basin such as Kula Krai and Tanah Merah. In addition, construction of new storage reservoirs, either elevated or ground reservoirs, particularly in the lower part of the basin is required, since most of the paddy fields are located in this region.

This study focuses on the quantification of hydro-meteorological droughts, the response rate analysis of SSI to SPI and the impact of the large-scale atmospheric circulation on the dry periods in the Kelantan River Basin (KRB) between 1985 and 2020. The main findings of the study are summarized as follows:

• KRB had experienced extremely dry events vary across the stations, but most of the stations were affected in 1986, 1987, 1989, 1990, 1992, 1997–1998, 2015–2016, and 2020. From January to March, there was a noticeable increasing trend in SPI-3, and some stations like Pej Hai Jajahan Machang, Ulu Sekor, Kuala Krai, and Kota Bharu showed a significant change at 95% confidence level. Furthermore, from April to December, both increasing and decreasing trends were seen. Most of the stations in the upper and middle parts of the basin, including Balai Polis Bertam, Pos WIas, Kuala Krai, and Kampung Aring, had a significant decreasing trend of SPI-3, especially from September to November, threatening the water supply of the KRB.

• The majority of the stations in the middle and upper parts of the KRB experienced more intense droughts than those in the lower part. The response rates of SSI-3 to SPI-3 were higher in December–February and March–May than in June–August and September–November, ranging from 50 to 64%. The moderate response rate may be contributed by the extraction of groundwater as the main freshwater supply in the KRB, where part of the rainwater has recharged the shallow groundwater.

• In general, ENSO played a more crucial role than IOD and MJO on the hydro-meteorological droughts formation in the KRB, especially from February to June. At a 95% confidence level, IOD was also correlated with the SPI-3 in March and SSI-3 in March and April. Hence, more attention should be paid by the local authorities and water supply company during the El Nino years.

• In the KRB, the paddy field was most negatively impacted during the dry spells due to the low water supplies for irrigation. Hence, more water treatment plants should be constructed in the areas that are less impacted by intense drought, i.e., in the middle part of the basin and along the Sungai Lebir (Figure 7(d)). In addition, since most of the paddy fields are in the lower part of the basin, building new storage reservoirs are essential to reduce the impact of drought on the agricultural sector.

Future studies should look at ways to enhance SPI drought monitoring, such as application of satellite products, development of a new satellite-based drought index and hydrological modelling (Li & Zhang 2008; Liu et al. 2020). In addition, recent studies have shown an interest in forecasting evaporation and soil moisture (Zhang et al. 2019; Zhao et al. 2022), which might be included in drought indices for even more accurate drought monitoring in Kelantan.

This project was supported by the Ministry of Higher Education Malaysia Long-term Research Grant Scheme Project 2 (grant no. LRGS/1/2020/UKM-USM/01/6/2), which is under the LRGS/1/2020/UKM/01/6 program.

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

The authors declare there is no conflict.