Abstract

Severe droughts in the year 1998 and 2014 in Sarawak due to the strong El Nino has impacted the water supply and irrigated agriculture. In this study, the Standardized Precipitation Index (SPI) was used for drought identification and monitoring in Sarawak River Basin. Using monthly precipitation data between the year 1975 and 2016 for 15 rainfall stations in the basin, the drought index values were obtained for the time scale of three, six and nine months. Rainfall trend for the years in study was also assessed using the Mann–Kendall test and Sen's slope estimator and compared with the drought index. Findings showed that generally there was a decreasing trend for the SPI values for the three time scales, indicating a higher tendency of increased drought event throughout the basin. Furthermore, it was observed that there was an increase in the numbers of dry months in the recent decade for most of the rainfall stations as compared to the previous 30 to 40 years, which could be due to climate change. Findings from this study are valuable for the planning and formulating of drought strategies to reduce and mitigate the adverse effects of drought.

INTRODUCTION

Drought is a natural hazard with large impacts on regular human activities, reduction of crop production and water shortages. Drought is linked to precipitation intensity, amount of precipitation occurrences and time scale between two wet seasons. The severe drought in 1998 was connected to the strong El Nino Southern Oscillation event, which affected millions of residents in Sarawak, Malaysia, caused high global temperatures, disrupted water supply in regional areas, caused forest fires and impacted irrigated agriculture. According to the World Meteorological Organization (WMO), six drought periods in Sarawak were during the periods 1982–1983, 1986–1988, 1991–1992, 1997–1998, 2009–2010 and 2014–2016 due to the strong El Nino.

There are several established scientific methods to identify and forecast drought occurrence, such as Standardized Precipitation Index (SPI), Palmer Severity Index, Crop Moisture Index and Reclamation Drought Index, which are commonly used to determine the drought indices. These drought index values incorporate thousands of data on rainfall, stream flow and other water resources indicators into an understandable large representation. Some indices are more suitable than others for certain uses, even though none of the main indices is essentially superior to the rest in all circumstances. Each of the indices works in a different way depending on the need that arises (Othman et al. 2016a).

Drought indices such as the SPI can be used to evaluate the impact of climate change on short- and medium-term drought in a region (Lee et al. 2017). This can be done either by determining the changes in frequency of extreme events that might accompany climate change (Loukas et al. 2008), analysis of severity, duration and intensity of time series SPI (Lee et al. 2017) and trends projection of SPI using possible future climate change scenarios (Huang et al. 2016; Osuch et al. 2016). For the case of Peninsular Malaysia, a study conducted by Yusof et al. (2012) using SPI based on 33 years of daily precipitation data for 69 stations has shown upward trend values for drought events, especially in eastern and western parts. This finding concurs with the latest study by Tang (2019), where increasing hot years are evident from the temperature surges under the influence of El Nino and the co-occurrence of dry spells and heavy rainfall within the same year is an emerging weather pattern in Malaysia.

This study aims to understand the trend of annual monthly precipitation and whether climate change has increased the frequency of drought events in recent years in Sarawak River Basin. The trend of annual monthly precipitation for the selected rainfall stations in Sarawak River Basin was investigated by using Mann–Kendall (MK) test and Sen's slope estimator and compares the outcome with drought index. The SPI was chosen as the tool to analyse and monitor drought events in this study. This study will provide insights into drought assessment and also the impact from climate change for Sarawak River Basin which is lacking in the existing literature.

METHODOLOGY

The location of the selected rainfall stations in this study are as shown in Figure 1. Sarawak River Basin is one of the major river basins that is situated at the southern part of Sarawak, Malaysia, with a catchment area of approximately 2,456 km2 and river length approximately 120 km. According to Hii et al. (2011), Sarawak River Basin mainly experiences two main monsoon seasons: northeast monsoon season (November–March) whereby the wet season is recorded and southwest monsoon season (June–September) whereby dry months are recorded. The climate of the river basin is classified as tropical rain forest that consists of high temperature and high annual total precipitation of about 3,830 mm (Abdillah et al. 2013). An earlier preliminary study by Bong et al. (2009) has shown that, generally, the mean annual rainfall, annual mean temperature and annual mean daily evaporation rate for Sarawak River Basin have an upward trend for the past three to four decades. However, limited studies on drought were found in the literature for the basin.

Figure 1

Location of selected rainfall stations in Sarawak River Basin (DID 2019).

Figure 1

Location of selected rainfall stations in Sarawak River Basin (DID 2019).

Rainfall data

Rainfall data and periods obtained for this study are shown in Table 1. The monthly rainfall data were collected from Drainage and Irrigation Department (DID), Sarawak. The selection of rainfall stations was based on sufficient rainfall record that was at least 30 years of data (in this case between the years 1975 and 2016). There are only 15 rainfall stations out of a total of 49 which meet the criteria as outlined, and these were used in this study.

Table 1

Details for the selected rainfall stations in Sarawak River Basin

Station ID Station name Location
 
Data period 
Latitude (°N) Longitude (°E) 
1102019 Padawan 01°09°48 110°15°19 1975–2016 
1201076 Semban 01°13°52 110°10°58 1975–2016 
1301074 Krokong 01°21°55 110°06°52 1980–2016 
1302078 Kampung Git 01°21°22 110°15°47 1986–2016 
1401005 Bau 01°25°06 110°08°58 1975–2016 
1402001 Siniawan W/W 01°26°46 110°12°37 1984–2016 
1402047 Batu Kitang 01°27°01 110°16°55 1980–2016 
1403001 Kuching Airport 01°29°27 110°20°57 1975–2016 
1502001 Sebubut 01°34°49 110°13°03 1980–2016 
1502026 Matang 01°34°39 110°12°29 1981–2016 
1503083 Kuching Third Mile 01°31°46 110°20°32 1981–2016 
1601001 Sungai Rayu 01°36°48 110°08°49 1986–2016 
1601003 Sungai China 01°36°47 110°11°35 1981–2016 
1603058 Rampangi 01°40°43 110°20°01 1985–2016 
1704013 Telok Assam 01°43°04 110°16°29 1981–2015 
Station ID Station name Location
 
Data period 
Latitude (°N) Longitude (°E) 
1102019 Padawan 01°09°48 110°15°19 1975–2016 
1201076 Semban 01°13°52 110°10°58 1975–2016 
1301074 Krokong 01°21°55 110°06°52 1980–2016 
1302078 Kampung Git 01°21°22 110°15°47 1986–2016 
1401005 Bau 01°25°06 110°08°58 1975–2016 
1402001 Siniawan W/W 01°26°46 110°12°37 1984–2016 
1402047 Batu Kitang 01°27°01 110°16°55 1980–2016 
1403001 Kuching Airport 01°29°27 110°20°57 1975–2016 
1502001 Sebubut 01°34°49 110°13°03 1980–2016 
1502026 Matang 01°34°39 110°12°29 1981–2016 
1503083 Kuching Third Mile 01°31°46 110°20°32 1981–2016 
1601001 Sungai Rayu 01°36°48 110°08°49 1986–2016 
1601003 Sungai China 01°36°47 110°11°35 1981–2016 
1603058 Rampangi 01°40°43 110°20°01 1985–2016 
1704013 Telok Assam 01°43°04 110°16°29 1981–2015 

Estimation of missing data

Estimation of missing data up to the year 2016 were made in this study to preserve the continuity of monthly precipitation data for SPI calculation and data consistency check to increase the accuracy of the SPI results. The percentages of missing data were calculated for all the selected precipitation stations. Only the precipitation stations with missing data less than 10% within the study period were selected for SPI analysis. Normal ratio method was applied to compute the missing value of the selected station by selecting the rainfall data of the three closest neighbouring stations.

Mann–Kendall (MK) test

The annual monthly rainfall trend for this study was determined by using two non-parametric tests, namely, Mann–Kendall (MK) test and Sen's slope estimator. The MK test and Sen's slope estimator have been commonly used to analyse trends for extreme rainfall events, such as the previous work done by Othman et al. (2016b) for the Pahang and Kelantan river basins in Malaysia. The MK test (Mann 1945; Kendall 1975; Gilbert 1987) is a non-parametric (distribution-free) numerical test usually proposed to investigate the trend within hydrological time scales. The function of the MK test is to statistically evaluate whether there is a monotonic increasing or decreasing trend of the variable of interest over time.

The Mann–Kendall statistic, S, is incremented by 1 if data value from a later time period is higher than data value sampled from an earlier time period. If no trend presents at first, the S value is assumed to be 0 and is increased by 1 if the later time period data are higher than the earlier time period. The S value decreases by 1 if the earlier time period data are higher than later time period data. After the net result of all the increments and decrements of data have been computed, the final S value will be produced. The S value will be a positive value if the trend present is an upward trend and a downward trend will be a negative value.

Hypothesis testing is applied to determine the significance of the trend. For the test interpretation, the null hypothesis (Ho) indicates that there is no significant trend in the data series while the alternative hypothesis (Ha) indicates a significant trend present in the data series. If the Z value is smaller than the selected significance level or confidence level, the null hypothesis is accepted.

Sen's slope estimator

Sen's slope estimator is a non-parametric test applied to estimate the magnitude of the trend in time series data (Mayowa et al. 2015). The true slope (change per unit time) can be approximated by conducting a straightforward non-parametric procedure developed by Sen (1968) if a linear trend is shown in a time series.

Sen's slope estimator is tested by a probability value (p-value) two-tailed test at a certain confidence level. A positive magnitude value shows increasing trend with time while negative magnitude value shows decreasing trend with time. Sen's slope estimator is commonly used with the MK test and the statistic S value of the MK test is consistent with Sen's slope value. If the MK test statistic presents an upward trend, a similar positive slope magnitude can be obtained.

SPI methodology using Gamma probability distribution

SPI is a simple index and straight forward to be computed and statistically related depending on the probability of rainfall for any time period. SPI can help to assess drought severity for any location and any period because of its normal distribution (McKee et al. 1993) and the frequencies are consistent. According to Hayes et al. (2011), the WMO suggested SPI as the main meteorological drought index that countries are supposed to use to monitor and investigate drought situations.

McKee et al. (1993) proposed the categorization system for the SPI value which is classified into seven categories, as in Table 2, to identify drought severity. A drought occurs at any time, as the SPI is continuously negative until it achieves a value of −1.0 or more. Drought ends when the SPI value becomes positive.

Table 2

SPI classification (McKee et al. 1993)

SPI value Class Cumulative probability Probability of event (%) 
SPI ≥ 2.00 Extreme wet 0.977–1.000 2.3% 
1.50 ≤ SPI < 2.00 Very wet 0.933–0.977 4.4% 
1.00 ≤ SPI < 1.50 Moderately wet 0.841–0.933 9.2% 
−1.00 ≤ SPI < 1.00 Near normal 0.159–0.841 68.2% 
−1.50 ≤ SPI < −1.00 Moderate dry 0.067–0.159 9.2% 
−2.00 ≤ SPI < −1.50 Severe dry 0.023–0.067 4.4% 
SPI < −2.00 Extreme dry 0.000–0.023 2.3% 
SPI value Class Cumulative probability Probability of event (%) 
SPI ≥ 2.00 Extreme wet 0.977–1.000 2.3% 
1.50 ≤ SPI < 2.00 Very wet 0.933–0.977 4.4% 
1.00 ≤ SPI < 1.50 Moderately wet 0.841–0.933 9.2% 
−1.00 ≤ SPI < 1.00 Near normal 0.159–0.841 68.2% 
−1.50 ≤ SPI < −1.00 Moderate dry 0.067–0.159 9.2% 
−2.00 ≤ SPI < −1.50 Severe dry 0.023–0.067 4.4% 
SPI < −2.00 Extreme dry 0.000–0.023 2.3% 

McKee et al. (1993) recommended using the Gamma probability distribution in SPI calculation and it has been used in Mishra & Desai (2006), Belayneh & Adamowski (2013) and Lloyd-Hughes & Saunders (2002). For this study, the SPI values were calculated by using SPI software program (SPI_SL_6.exe) which can be downloaded for free from the National Drought Mitigation Center, University of Nebraska Lincoln site (http://drought.unl.edu/MonitoringTools/DownloadableSPIProgram.aspx).

Theoretically, the SPI represents the number of standard deviations, above or below that an event is from the mean. The SPI will have a standard normal distribution with an expected value of 0 and a variance of 1 during the base period when the gamma parameters are estimated. After the SPI values are determined for the different time scales, the frequency/number of occurrences for drought for each decade can be determined by using the SPI value below −1.0 as the threshold value for drought condition (Loukas et al. 2008; Lee et al. 2017).

RESULTS AND DISCUSSION

Rainfall trend analysis

The missing values for stations Semban, Kampung Git and Sungai Rayu were determined to be more than 10% and were therefore excluded from this study. The estimation for missing values was done by using normal ratio method for the other 12 rainfall stations by interpolating the data of the selected station with the three neighbouring stations to obtain a complete dataset. Addinsoft's XLSTAT 2018 software package was applied for the non-parametric tests. The MK test and Sen's slope estimator with 95% confidence level were used to investigate the trend significance and magnitude, respectively. Table 3 shows the results of the trend analysis.

Table 3

Trend analysis results

Station ID Station name Total monthly rainfall
 
S p-value (two-tailed) Sen's slope (mm/year) 
1102019 Padawan 874 0.816 0.009 
1301074 Krokong −2,214 0.477 −0.042 
1401005 Bau −1,228 0.744 −0.015 
1402001 Siniawan W/W −2,555 0.330 −0.057 
1402047 Batu Kitang −5,495 0.078 −0.094 
1403001 Kuching Airport 511 0.892 0.006 
1502001 Sebubut −2,029 0.513 −0.035 
1502026 Matang −4,816 0.107 −0.108 
1503083 Kuching Third Mile 3,711 0.214 0.069 
1601001 Sungai Rayu −3,068 0.220 −0.101 
1603058 Rampangi 383 0.873 0.011 
1704013 Telok Assam −2,086 0.466 −0.048 
Station ID Station name Total monthly rainfall
 
S p-value (two-tailed) Sen's slope (mm/year) 
1102019 Padawan 874 0.816 0.009 
1301074 Krokong −2,214 0.477 −0.042 
1401005 Bau −1,228 0.744 −0.015 
1402001 Siniawan W/W −2,555 0.330 −0.057 
1402047 Batu Kitang −5,495 0.078 −0.094 
1403001 Kuching Airport 511 0.892 0.006 
1502001 Sebubut −2,029 0.513 −0.035 
1502026 Matang −4,816 0.107 −0.108 
1503083 Kuching Third Mile 3,711 0.214 0.069 
1601001 Sungai Rayu −3,068 0.220 −0.101 
1603058 Rampangi 383 0.873 0.011 
1704013 Telok Assam −2,086 0.466 −0.048 

No significant trend was shown in the series among the 12 stations with p-value more than α value (0.05). Eight stations presented decreasing trend and four stations presented increasing trend in the series. The analysis results revealed that most of the rainfall stations had a decreasing trend in annual monthly rainfall. Hence, there was a symptom of rainfall decreases over this minimum 30-years study period. Reduction of rainfall trend might lead to severe drought in the future.

Variability of SPI time series

Drought studies are normally focused on the three- to nine-month time scales. Different time scales of the SPI presented differences in magnitude and duration of droughts. For short- and medium-term time scale, three-month SPI was used to analyse moisture conditions and hence available to provide a seasonal estimation of precipitation. The three-month SPI compares the rainfall for a specific three-month period with the total rainfalls from the same three-month period for all the years chosen in the historical record. The six-month SPI compares the rainfall for the selected period with the same six-month period over the historical report. Six-month SPI specifies medium-term trends in rainfall and is very effective in showing the precipitation over distinct seasons. Nine-month SPI is considered a hydrological drought index and becomes useful for monitoring the surface water resources. A downward trend of SPI value indicates a higher tendency of increased drought occurrence throughout the basin.

Most of the rainfall stations with the same time scales presented a downward trendline in the series. For the three-month SPI time scale, there were three rainfall stations showing an upward trendline but only two rainfall stations for six-month SPI and one rainfall station for nine-month SPI. The tendency of fewer but longer droughts was found in the six- and nine-month SPI averaging periods as compared to three-month SPI. This concurs with the findings of Naumann et al. (2018), where global median drought length is projected to slowly increase with global warming and climate change. A downward trend of SPI values indicates a higher possibility of increased drought happening throughout the basin, as shown in Figure 2 for Matang station. Similar trends were observed by Yusof et al. (2012) for daily rainfall data between the periods of November 1975 and October 2008 where most of Peninsular Malaysia experiences drier drought event. Table 4 summarizes the SPI values for the selected rainfall stations in Sarawak River Basin with the year the peak intensity of drought occurs for each of the time scales.

Table 4

Summary of drought occurrences for the selected rainfall stations in Sarawak River Basin

Rainfall station SPI interval (months) Duration (months) Mean intensity Peak intensity
 
SPI value Year 
Padawan 504 −0.01 −2.86 1982 
504 −0.01 −2.63 1975 
504 −0.01 −2.85 1975 
Krokong 444 −0.01 −3.21 2014 
444 −0.01 −3.1 2014 
444 −0.01 −2.91 2014 
Bau 504 −0.02 −2.81 1996 
504 −0.02 −2.64 1996 
504 −0.02 −3.12 1996 
SiniawanW/W 384 −0.01 −3.16 2014 
384 −0.01 −3.15 1998 
384 −0.01 −3.31 1998 
Batu Kitang 444 −0.02 −2.72 2016 
444 −0.01 −3.01 2006 
444 −0.01 −2.47 1991 
Kuching Airport 504 −0.01 −3.22 1997 
504 −0.01 −3.6 1997 
504 −0.01 −3.0 1998 
Sebubut 444 −0.01 −2.32 1983 
444 −0.02 −2.44 1991 
444 −0.01 −2.31 2006 
Matang 432 −0.01 −2.42 2016 
432 −0.01 −2.94 1991 
432 −0.01 −2.83 1991 
Kuching Third Mile 432 −0.01 −2.96 1996 
432 −0.01 −3.05 1998 
432 −0.01 −3.24 1998 
Rampangi 372 −0.01 −3.61 1997 
372 −0.01 −3.13 1997 
372 −0.01 −3.14 1997 
Sungai Rayu 384 −0.01 −2.51 2001 
384 −0.02 −2.25 2009 
384 −0.01 −2.18 2001 
Telok Assam 432 −0.01 −2.87 1997 
432 −0.01 −3.08 1998 
432 −0.01 −3.53 1998 
Rainfall station SPI interval (months) Duration (months) Mean intensity Peak intensity
 
SPI value Year 
Padawan 504 −0.01 −2.86 1982 
504 −0.01 −2.63 1975 
504 −0.01 −2.85 1975 
Krokong 444 −0.01 −3.21 2014 
444 −0.01 −3.1 2014 
444 −0.01 −2.91 2014 
Bau 504 −0.02 −2.81 1996 
504 −0.02 −2.64 1996 
504 −0.02 −3.12 1996 
SiniawanW/W 384 −0.01 −3.16 2014 
384 −0.01 −3.15 1998 
384 −0.01 −3.31 1998 
Batu Kitang 444 −0.02 −2.72 2016 
444 −0.01 −3.01 2006 
444 −0.01 −2.47 1991 
Kuching Airport 504 −0.01 −3.22 1997 
504 −0.01 −3.6 1997 
504 −0.01 −3.0 1998 
Sebubut 444 −0.01 −2.32 1983 
444 −0.02 −2.44 1991 
444 −0.01 −2.31 2006 
Matang 432 −0.01 −2.42 2016 
432 −0.01 −2.94 1991 
432 −0.01 −2.83 1991 
Kuching Third Mile 432 −0.01 −2.96 1996 
432 −0.01 −3.05 1998 
432 −0.01 −3.24 1998 
Rampangi 372 −0.01 −3.61 1997 
372 −0.01 −3.13 1997 
372 −0.01 −3.14 1997 
Sungai Rayu 384 −0.01 −2.51 2001 
384 −0.02 −2.25 2009 
384 −0.01 −2.18 2001 
Telok Assam 432 −0.01 −2.87 1997 
432 −0.01 −3.08 1998 
432 −0.01 −3.53 1998 
Figure 2

SPI for Matang station: (a) three months; (b) six months; (c) nine months.

Figure 2

SPI for Matang station: (a) three months; (b) six months; (c) nine months.

Application of the SPI in drought monitoring

Water supply shortage happened in 2014 in Malaysia due to the strong El Nino phenomenon causing extensive drought and absence of rains that affected agriculture and food production. For short- and medium-term time scale, three-month SPI was used to analyse moisture conditions and hence available to provide a seasonal estimation of precipitation. In the year 2014, the driest region in Sarawak River Basin was around Siniawan W/W rainfall station with the lowest SPI value of −3.21 and the wettest region was Padawan with the highest SPI value of 1.74, as shown in Table 5. SPI values of more than 1.0, mostly from August to October which is considered as wet season, indicates heavy rains and recurrent localized floods occur while the SPI values less than −1.0 from January to April were classified as dry. The most suitable period for growing crops is from December to March with SPI values between 1 and −1 as the growing season begins. As mentioned in Table 2, SPI values of −1 to 1 are considered near normal weather condition, that is, not too wet or too dry and suitable for planting. The most suitable harvesting season is from April to June with SPI values of less than 1. The SPI values continue to drop to −3 in July and illustrate that the weather is extreme dry and drought may occur. Tables 6 and 7 show the monthly values for six-month and nine-month SPI, respectively, for Sarawak River Basin in the year 2014. As the time scale increases, the SPI responds more slowly to short-term variation of rainfall and the cycles of negative SPI values become more visible for the year 2014 due to the El Nino phenomenon. However, the severity of the drought (based on the SPI values) seems to reduce slightly as the time scale increases.

Table 5

The monthly values for three-month SPI in 2014

 
 
Table 6

The monthly values for six-month SPI in 2014

 
 
Table 7

The monthly values for nine-month SPI in 2014

 
 

Impact of climate change on drought frequency

Table 8 shows the total number of dry months for different decades for each of the time scales in the current study. Generally, most stations in Sarawak River Basin showed that the decade between the years 1997 and 2006 has the highest numbers of dry months with SPI values less than −1 (except for Padawan) as compared to the other decades. This higher number of dry months during the period 1997–2006 could be due to the super El Nino phenomenon that happened in 1997–1998. However, looking at a longer period, the most recent decade period of 2007–2016 tends to have higher numbers of dry months for most of the stations (except for Bau, Matang, Teluk Assam and Kuching Third Mile) as compared to the period 1977–1986 and 1987–1996. This could be due to the impact of climate change, although more detailed study is needed to ascertain this hypothesis.

Table 8

Total numbers of dry months (SPI < −1) for different decades

Sub-basin Period SPI 3 SPI 6 SPI 9 
Padawan 1977–1986 17 19 15 
1987–1996 12 13 13 
1997–2006 18 17 18 
2007–2016 13 23 26 
Bau 1977–1986 15 18 25 
1987–1996 10 11 11 
1997–2006 20 30 29 
2007–2016 15 15 15 
Krokong 1987–1996 13 12 17 
1997–2006 29 31 30 
2007–2016 16 14 20 
Kuching Airport 1977–1986 17 17 15 
1987–1996 12 11 
1997–2006 24 22 28 
2007–2016 12 17 20 
Siniawan 1987–1996 14 17 13 
1997–2006 22 26 31 
2007–2016 20 17 19 
Matang 1987–1996 23 18 18 
1997–2006 30 27 27 
2007–2016 14 12 10 
Sebubut 1987–1996 13 
1997–2006 31 28 29 
2007–2016 16 17 14 
Batu Kitang 1987–1996 17 14 16 
1997–2006 19 17 19 
2007–2016 26 23 23 
Sungai Rayu 1987–1996 13 11 13 
1997–2006 18 21 20 
2007–2016 20 28 23 
Rampangi 1987–1996 16 
1997–2006 21 13 13 
2007–2016 16 10 
Teluk Assam 1987–1996 20 20 21 
1997–2006 17 21 23 
2007–2016 13 16 12 
Kuching Third Mile 1987–1996 18 17 15 
1997–2006 21 25 28 
2007–2016 
Sub-basin Period SPI 3 SPI 6 SPI 9 
Padawan 1977–1986 17 19 15 
1987–1996 12 13 13 
1997–2006 18 17 18 
2007–2016 13 23 26 
Bau 1977–1986 15 18 25 
1987–1996 10 11 11 
1997–2006 20 30 29 
2007–2016 15 15 15 
Krokong 1987–1996 13 12 17 
1997–2006 29 31 30 
2007–2016 16 14 20 
Kuching Airport 1977–1986 17 17 15 
1987–1996 12 11 
1997–2006 24 22 28 
2007–2016 12 17 20 
Siniawan 1987–1996 14 17 13 
1997–2006 22 26 31 
2007–2016 20 17 19 
Matang 1987–1996 23 18 18 
1997–2006 30 27 27 
2007–2016 14 12 10 
Sebubut 1987–1996 13 
1997–2006 31 28 29 
2007–2016 16 17 14 
Batu Kitang 1987–1996 17 14 16 
1997–2006 19 17 19 
2007–2016 26 23 23 
Sungai Rayu 1987–1996 13 11 13 
1997–2006 18 21 20 
2007–2016 20 28 23 
Rampangi 1987–1996 16 
1997–2006 21 13 13 
2007–2016 16 10 
Teluk Assam 1987–1996 20 20 21 
1997–2006 17 21 23 
2007–2016 13 16 12 
Kuching Third Mile 1987–1996 18 17 15 
1997–2006 21 25 28 
2007–2016 

CONCLUSIONS

The analysis results generally showed decreasing trend for the SPI values for the three time scales which indicates a higher tendency of increased drought events throughout the basin. Also, an analysis on the numbers of dry months for the past decades has shown there is a tendency of increased drought events in the recent decade with more prolonged periods. Hence, it is predicted that drought will hit harder and for more prolonged periods in the future for the Sarawak River Basin. These findings indicated that climate change could affect the drought severity and subsequently the planning of water resources projects and drought management in the basin. Future work involving spatial mapping of SPI is suggested for the investigation of temporal and spatial variability of drought occurrences in the basin and their connection with possible hazards such as forest fires and floods.

ACKNOWLEDGEMENTS

The authors would like to acknowledge and thank Universiti Malaysia Sarawak for the use of the computing facilities and Ministry of Higher Education (MOHE) Malaysia for the financial support under the Fundamental Research Grant Scheme (Grant No. F02/FRGS/1618/2017). The authors also would like to express their deepest appreciation to Ms. Tay Siew Voon of Hydrology and Water Resources Branch, Department of Irrigation and Drainage (DID) Sarawak for providing the authors with constructive suggestions and data for the project.

REFERENCES

REFERENCES
Abdillah
N.
,
Teo
F. Y.
,
Jusoh
A. M.
,
Fauzi
M. F.
,
Hassan
A. M. M.
&
Darus
A.
2013
Model assessment of water quality in Sarawak River, Malaysia
. In:
Proceedings of 35th IAHR World Congress
,
Cheng Du, China
, pp.
1
9
.
Belayneh
A.
&
Adamowski
J.
2013
Drought Forecasting using new machine learning method
.
Journal of Water and Land Development
18
(
9
),
3
12
.
Bong
C. H. J.
,
Ting
S. Y.
,
Bustami
R. A.
&
Putuhena
F. J.
2009
Impact of climate change and its variability on the rainfall pattern in Sarawak River Basin
. In:
Proceedings of International Conference on Water Resoures (ICWR 2009)
,
Langkawi, Kedah, Malaysia
.
DID
2019
Resource Centre – Hydrology Stations At Sg Sarawak Basin
. .
Gilbert
R. O.
1987
Statistical Methods for Environmental Pollution Monitoring
.
John Wiley & Sons, Inc
,
New York
,
USA
.
Hayes
M.
,
Svoboda
M.
,
Wall
N.
&
Widhalm
M.
2011
The Lincoln Declaration on drought indices: universal meteological drought index recommended
.
Bulletin of the American Meteorogical Society
92
(
4
),
485
488
.
Hii
C. P.
,
Putuhena
F. J.
&
Said
S.
2011
Floodplain and hydraulic infrastructure system modelling for adaptive flood management
.
The Journal of the Institution of Engineers Malaysia
72
(
2
),
276
285
.
Huang
Y. F.
,
Ang
J. T.
,
Tiong
Y. J.
,
Mirzaei
M.
&
Mat Amin
M. Z.
2016
Drought forecasting using SPI and EDI under RCP-8.5 climate change scenarios for Langat River Basin, Malaysia
.
Procedia Engineering
154
,
710
717
.
Kendall
M. G.
1975
Rank Correlation Methods
,
4th edn
.
Charles Griffin
,
London
,
UK
.
Lloyd-Hughes
B.
&
Saunders
M. A.
2002
A drought climatology for Europe
.
International Journal of Climatology
22
(
13
),
1571
1592
.
Loukas
A.
,
Vasiliades
L.
&
Tzabiras
J.
2008
Climate change effects on drought severity
.
Advances in Geosciences
17
,
23
29
.
Mann
H. B.
1945
Nonparametric tests against trend
.
Econometrica
13
(
3
),
163
171
.
Mayowa
O. O.
,
Pour
H. S.
,
Shahid
S.
,
Mohsenipour
M.
,
Harun
S.
,
Heryansyah
A.
&
Ismail
T.
2015
Trends in rainfall and rainfall-related extremes in the east coast of peninsular Malaysia
.
Indian Academy of Sciences
124
(
8
),
1609
1622
.
McKee
T. B.
,
Doesken
N. J.
&
Kleist
J.
1993
The relationship of drought frequency and duration to time scale
. In:
Proceedings of the Eighth Conference on Applied Climatology
,
Anaheim, CA
, pp.
17
22
.
Mishra
A. K.
&
Desai
V. R.
2006
Drought forecasting using feed forward recursive neural network
.
Ecological Modelling
198
(
1
),
127
138
.
Naumann
G.
,
Alfieri
L.
,
Wyser
K.
,
Mentaschi
L.
,
Betts
R. A.
,
Carrao
H.
,
Spinoni
J.
,
Vogt
J.
&
Feyen
L.
2018
Global changes in drought conditions under different levels of warming
.
Geophysical Research Letters
45
,
3285
3296
.
Osuch
M.
,
Romanowicz
R. J.
,
Lawrence
D.
&
Wong
W. K.
2016
Trends in projections of standardized precipitation indices in a future climate in Poland
.
Hydrology and Earth System Sciences
20
,
1947
1969
.
Othman
M.
,
Ash'aari
Z. H.
,
Muharam
F. M.
,
Sulaiman
W. N. A.
,
Hamisan
H.
,
Mohamad
N. D.
&
Othman
N. H.
2016a
Assessment of drought impacts on vegetation health: a case study in Kedah
.
IOP Conference Series: Earth and Environmental Science
37
(
1
),
12
72
.
Othman
M. A.
,
Zakaria
N. A.
,
Ab. Ghani
A.
,
Chang
C. K.
&
Chan
N. W.
2016b
Analysis of trends of extreme rainfall events using Mann Kendall Test: a case study in Pahang and Kelantan river basins
.
Jurnal Teknologi
78
,
63
69
.
Sen
P. K.
1968
Estimates of the regression coefficient based on Kendall's tau
.
Journal of the American Statistical Association
63
(
324
),
1379
1389
.
Yusof
F.
,
Foo
H.-M.
,
Suhaila
S.
&
Kong
C.-Y.
2012
Trend analysis for drought event in Peninsular Malaysia
.
Jurnal Teknologi
57
(
1
),
211
218
.