Spatiotemporal Variability of Seasonal Precipitation

in Central Dry Zone, Myanmar

Yin Yin Htwe1, Thet Zin Htoo1,2
1Department of Civil Engineering, Mandalay Technological University, Mandalay, 05072, Myanmar;
2
Department of Civil Engineering, Pandit Deendayal Energy University, Gandhinagar, 382426, Gujarat, India.
[email protected], [email protected]

Abstract – This study investigates seasonal precipitation cycle. Precipitation is also one of the most important
trends across ten townships in the Central Dry Zone (CDZ) climatic variables that has a direct and indirect impact of
of Myanmar using the non-parametric Mann-Kendall (MK) agricultural production and ecosystem health. Changes in
test and Sen’s slope estimator. A 31-year record of rainfall intensity, quantity, and spatiotemporal patterns
historical precipitation data (1992–2022) was utilized to often lead to extreme events such as droughts and floods.
assess seasonal variations. The objective was to examine These changes may require adjustments to current strategies
the spatiotemporal variability of rainfall and its of planning and management of water resources. Analysis
implications for water resource management and of the spatial distribution and temporal trends of
agricultural planning in this climatically vulnerable region. precipitation is crucial for water resource management,
The results revealed significant seasonal and spatial agricultural productivity and climate change mitigation
variability in precipitation trends. During the summer [1],[2],[3].
season, most stations exhibited either a slight positive trend Myanmar is one of the country’s most vulnerable
or no statistically significant change, indicating relatively to the impacts of climate change, with its agriculture-based
stable conditions. In contrast, the rainy season showed a economy and communities highly sensitive to fluctuations
general increasing trend across most townships, with the in weather and water availability [4], [5]. In recent years,
most pronounced increases observed in Naung-U, the monsoon weather pattern, Myanmar's primary source of
Yemathin, and Magway. Conversely, Shwebo and Monywa rainfall has become increasingly unpredictable in terms of
experienced decreasing trends, suggesting localized onset, duration, and intensity, leading to both droughts and
declines in monsoonal rainfall. No significant trends were floods across different regions [1], [7]. Analyzing and
detected during the winter season across all stations. understanding the spatiotemporal characteristics of
Finally, QGIS software was used to develop seasonal trend precipitation in Myanmar is essential for addressing water-
maps of precipitation, providing a visual representation of
related challenges such as water scarcity, declining
the spatial distribution of trends. These findings underscore
agricultural productivity, and increasing competition over
the complex nature of precipitation dynamics in the CDZ
water resources [8]. Moreover, human activities such as
and highlight the need for localized climate adaptation
deforestation, land-use change, and the expansion of
strategies to enhance sustainable agriculture and water
irrigated agriculture further disrupt the natural water cycle,
resource management in the face of ongoing climate
amplifying the impacts of climate variability and
change.
hydrological extremes [9].
The Central Dry Zone (CDZ) of Myanmar is
Keywords: Central dry zone, Precipitation, Non- increasingly affected by climate variability, with
parametric approach, Seasonal Variability, precipitation patterns becoming more erratic and
QGIS software
unpredictable. Given that agriculture and water supply in
the CDZ are heavily reliant on seasonal rainfall, changes in
precipitation directly influence crop yields, water
INTRODUCTION availability, and rural livelihoods. Despite the critical role
Climate change occurs when changes in Earth’s of precipitation, there is a lack of detailed studies on the
climate system result in new weather patterns that last for at spatial and temporal variability of rainfall in the region.
least a few decades, and may be for millions of years. Without reliable information and analysis, local authorities
Climate change is likely to affect hydrological cycle and farmers are unable to effectively plan for climate-
through precipitation, evapotranspiration and soil moisture resilient agricultural practices and water resource
etc. Global climate change has been an increasing challenge management. This gap presents a significant challenge to
to agricultural ecosystems, which will significantly affect achieving sustainable development and food security in one
the reference evapotranspiration. All climatic parameters as of Myanmar’s most vulnerable regions. Therefore, this
well as precipitation and evapotranspiration are influenced study was undertaken to analyze long-term precipitation
by climate change. Changes in the world’s climate have trends in the CDZ, providing essential data and insights
significant effect on water resources which affect the needed to support informed decision-making and promote
livelihood of people especially in arid and semi-arid regions sustainable development in one of Myanmar’s most
[1],[2]. Precipitation is a key component of the global water climate-sensitive regions.
OBJECTIVES significant vulnerability to climate variability. The
The objectives of this study are: locations of the study area are illustrated in Figure 1.
1. To analyze long-term precipitation trends in the
Central Dry Zone of Myanmar using non- Collection of Meteorological Data
parametric statistical methods. In this study, precipitation of 31 years record (1992-
2. To develop seasonal precipitation trend maps to 2002) are collected from the Department of Meteorology
visualize changes across the Central Dry Zone. and Hydrology, Upper Myanmar. To analyze the seasonal
variability of precipitation across this climate-sensitive
STUDY AREA area, ten meteorological stations have been selected based
The boundary of the central dry zone region of on their geographic distribution, data availability, and
Myanmar is generally defined between the North latitudes representation of key environmental conditions. These
19° 20′ to 22° 50′ and East longitudes 93°40′ to 96° 30′. The include five stations in Mandalay Region (Yamethin,
study area encompasses three major regions in Myanmar: Kyaukse, NaungU, Meilhtila, Myingyan), three in Sagaing
Mandalay Region, Sagaing Region, and Magway Region Region (Monywa, Shwebo, Sagaing), and three in Magway
which located within the Central Dry Zone (CDZ), a region Region (Magway, Minbu). The selected stations are
characterized by low rainfall, high temperatures, and presented in Figure 2.

Figure 1. Locations of the study area

METHODOLODY precipitation data to capture seasonal variations and

Trend detection of precipitation was conducted evaluate the statistical significance of detected trends.[5][6]
using two widely applied non-parametric statistical
methods: the Mann-Kendall (MK) test and Sen’s slope A. Mann-Kendall test
estimator. The MK test was used to assess trends in the data, The Mann-Kendall test is a non-parametric test
while Sen’s slope estimator quantified the rate of change used for identifying trends in time series data. This test is
over time. Both methods were applied to monthly carried out on monthly precipitation.
 Figure 2. Locations of the selected stations in central dry zone

Kendall’s statistic S is obtained using the B. Sen-Slope Estimator Test

following equations; Sen’s estimator of slope, which is a non -
n 1 parametric method, was used to develop linear models in
 sgn x  xk 
n
S j (1) this study. This test is also done based on monthly
k 1 j k 1 precipitation. This method offers many advantages that
 1, if x j  x k  0 have made it useful in analyzing atmospheric chemistry
 data. Missing values are allowed and the data need not
sgn  0, if x j  x k   0 (2) conform to any particular distribution.

 1, if x j  x k  0
 If a linear trend is present, the true slope (change
per unit time) can be estimated by using a simple non-
where, parametric procedure developed by Sen. The test procedure
S = Mann-Kendall’s statistic, is given below in detail. The slope estimates of N pairs of
xj, xk = data values at time j and k (j>k). data are first computed by
xj –xk
Qi = i=1,……., N (5)
g j–k
n n  12n  5   t i t i  12 t i  5 where,
j1
(3) Qi = slope estimates of the pairs of data
V
18 xj ,xk = data values at times j and k ( j > k)
where,
V = variance of S, The median of these N values of Qi is Sen’s
n = number of periods, estimator of slope. If there is only one datum in each time
ti =number of ties in group i. period, then N = n (n–1)/2 where n is the number of time
periods. If N is odd, then Sen’s estimator is computed by
 S  1 
 V 0 . 5 , for  S  0 Qmedian = Q N+1 (6)
 (4) 2
Z s   0 , for  S  0 If N is even, then Sen’s estimator becomes
 S  1 
 , for  S  0 Q   Q 
 V 0 .5  N 2    N  2  2  (7)
where, Q median 
2
Zs = Normal Standard Deviate
 The detected value of Qmedian is tested by a two- precipitation at the 90% confidence level, while Nyaung-U
sided test at the 100(1 –α) % confidence interval and true also demonstrates a statistically significant increasing trend
slope may be obtained by the non-parametric test. at the 95% confidence level. These results highlight
localized shifts in winter precipitation, particularly in
Nyaung-U, which may have implications for agricultural
RESULTS AND DISCUSSION planning and water resource management. The remaining
townships show no significant trend in winter precipitation
Seasonal precipitation trends across ten townships in the at any of the evaluated confidence levels, indicating
Central Dry Zone (CDZ) of Myanmar were analyzed using stability in those areas. Moreover, when examining monthly
31 years of monthly precipitation data (1992–2022). The precipitation trends during the winter season, only
non-parametric Mann-Kendall (MK) test was applied to Yamethin township shows a statistically significant upward
detect the presence and direction of trends, while Sen’s trend at the 90% confidence level. All other townships show
no significant trends at any of the selected confidence
slope estimator was used to quantify the rate of change.
levels, further emphasizing the localized nature of observed
Trend analysis was conducted on a seasonal basis; summer,
changes.
rainy, and winter to capture variations in rainfall patterns
across different times of the year. Table II. Mann-Kendall trend test for rainy season
level of confidence
A. Mann-Kendall trend test for seasonal analysis No station Z Statistics
The rank-based nonparametric Mann-Kendall method 90% 95% 99%
was applied to the long-term data to detect statistically 1 Shwebo z = -0.7603 No Trend No Trend No Trend
significant trends for 90%, 95% and 99% confidence level.
2 Monwya z = -0.8143 No Trend No Trend No Trend
In this test, the null hypothesis (H ) was that there has been
3 Sagaing z = -0.0734 No Trend No Trend No Trend
no trend in precipitation over time; the alternate hypothesis
(H ) was that there has been a trend (increasing or 4 Myingyan z = 1.5595 No Trend No Trend No Trend
decreasing) over time. If |𝑍| > Z / , (H ) is rejected and 5 Yemathin z = 1.8251 Increasing No Trend No Trend
a statistically significant trend exists in the hydrologic time 6 Kyaukse z = 1.1296 No Trend No Trend No Trend
series. The critical value of Z / for 90%,95% and 99% 7 Magway z = 1.4904 No Trend No Trend No Trend
confidence level are 1.64, 1.96 and 2.54 respectively. 8 Meikhtila z = 0.9028 No Trend No Trend No Trend
Z statistics is calculated using described equations and the 9 NaungU z = 2.2744 Increasing Increasing No Trend
results of the test for summer, rainy and winter season are 10 Minbu z = 1.3155 No Trend No Trend No Trend
presented in Table I, II and III respectively. And QGIS
software are used to develop map for seasonal variability Table III. Mann-Kendall trend test for winter season
with 90%, 95% and 99% confidence level are shown in
Figure 3, Figure 4 and Figure 5 respectively. level of confidence
No station Z Statistics
90% 95% 99%
Table I. Mann-Kendall trend test for summer season 1 Shwebo z = 0.1883 No Trend No Trend No Trend
level of confidence 2 Monwya z = -0.2514 No Trend No Trend No Trend
No station Z Statistics
90% 95% 99% 3 Sagaing z = -0.0743 No Trend No Trend No Trend
1 Shwebo z = 0.7112 No Trend No Trend No Trend 4 Myingyan z = 0.7384 No Trend No Trend No Trend
2 Monwya z = 0.9899 No Trend No Trend No Trend 5 Yemathin z = 1.9164 Increasing No Trend No Trend
3 Sagaing z = 0.4544 No Trend No Trend No Trend 6 Kyaukse z = 1.528 No Trend No Trend No Trend
4 Myingyan z = 0.2150 No Trend No Trend No Trend 7 Magway z = -0.8150 No Trend No Trend No Trend
5 Yemathin z = 0.3439 No Trend No Trend No Trend 8 Meikhtila z = 0.5835 No Trend No Trend No Trend
6 Kyaukse z = 1.5148 No Trend No Trend No Trend 9 NaungU z = 0.1454 No Trend No Trend No Trend
7 Magway z = 0.4293 No Trend No Trend No Trend 10 Minbu z = 0.1693 No Trend No Trend No Trend
8 Meikhtila z = 0.7664 No Trend No Trend No Trend
9 NaungU z = 0.1914 No Trend No Trend No Trend
B. Sen’s Slope estimator test for seasonal analysis
10 Minbu z = 0.424 No Trend No Trend No Trend
The seasonal precipitation trends across ten
According to the Mann-Kendall (MK) test for the meteorological stations in the Central Dry Zone were
summer season, the analysis indicates no statistically assessed using Sen’s slope estimator, which indicates both
significant trends in precipitation at any of the selected the direction and magnitude of change in precipitation over
confidence levels (90%, 95%, and 99%) across all time. Sen’s Slope test results for monthly precipitation in
townships, suggesting a relatively stable precipitation seasonal analysis at selected stations were presented in
pattern during this season. In contrast, the winter season Table IV. Map of variability for monthly precipitation are
reveals notable spatial variations. Both Yamethin and developed using QGIS sosftware and demostrated in Figure
Nyaung-U townships exhibit increasing trends in 6, Figure 7 and Figure 8.
 (a) 90 % confidence level (b) 95 % confidence level (c)99 % confidence level

Figure 3. Location of site with trend at 90%, 95% and 99% confidence level for monthly precipitation (summer season)

(a) 90 % confidence level (b) 95 % confidence level (c) 99 % confidence level

Figure 4. Location of site with trend at 90%, 95% and 99% confidence level for monthly precipitation (rainy season)

(a) 90 % confidence level (b) 95 % confidence level (c)99 % confidence level

Figure 5. Location of site with trend at 90%, 95% and 99% confidence level for monthly precipitation (winter season)
 Table IV. Sen’s Slope estimator test for seasonal analysis
Sen's slope (Qmedian)
No station
Summer Rainy Winter
1 Shwebo 0.0118 Positive -0.1745 Negative 0 No Trend
2 Monwya 0.0833 Positive -0.1785 Negative 0 No Trend
3 Sagaing 0.0108 Positive -0.0124 Negative 0 No Trend
4 Myingyan 0 No Trend 0.2500 Positive 0 No Trend
5 Yemathin 0 No Trend 0.2964 Positive 0 No Trend
6 Kyaukse 0.1382 Positive 0.1585 Positive 0 No Trend
7 Magway 0 No Trend 0.2849 Positive 0 No Trend
8 Meikhtila 0.0203 Positive 0.1937 Positive 0 No Trend
9 NaungU 0 No Trend 0.3872 Positive 0 No Trend
10 Minbu 0 No Trend 0.2068 Positive 0 No Trend

In the summer season, most stations showed either

a positive trend or no significant trend. Positive trends were
observed in Shwebo (0.0118), Monywa (0.0833), Sagaing
(0.0108), Kyaukse (0.1382), and Meikhtila (0.0203),
suggesting a gradual increase in precipitation. These
increases, although modest in magnitude, may reflect
localized variability in summer rainfall. On the other hand,
Myingyan, Yemathin, Magway, Naung-U, and Minbu
showed no significant trend, indicating stable summer
precipitation over the analysis period.During the rainy
season, most stations exhibited a positive trend, indicating
an increasing pattern in rainfall. The highest positive Sen’s
slope was recorded at Naung-U (0.3872), followed by
Yemathin (0.2964), Magway (0.2849), Myingyan (0.25),
and Minbu (0.2068). Other stations such as Meikhtila
(0.1937), Kyaukse (0.1585), and Monywa (–0.1785) also
showed notable trends.

Figure 7. Sen’s Slope trend map of stations with

increasing trends and decreasing trend in precipitation
during the rainy season

However, Monywa and Shwebo demonstrated negative

slopes (–0.1785 and –0.1745 respectively), suggesting a
decreasing rainfall trend in the rainy season for these
locations. These results indicate that while most areas are
experiencing increased monsoonal rainfall, a few stations
may be facing reduced rainfall, which could impact water
availability and crop productivity.The winter season shows
no significant trend across all ten stations, with all Sen’s
slope values being zero. This consistent result suggests that
winter precipitation patterns have remained stable over the
analysis period in the Central Dry Zone. The lack of trend
may reflect the generally low and variable nature of winter
Figure 6. Sen’s Slope trend map of stations with rainfall in this region, which typically contributes the least
increasing trends and no trend in precipitation during the to the annual precipitation total.
summer season.
 [2] FAO. (2016). Climate change and food security: risks and
responses. Food and Agriculture Organization of the United
Nations.
[3] WMO. (2012). International Glossary of Hydrology (3rd
ed.). World Meteorological Organization.
[4] UNDP (2012), Climate Change Vulnerability in Myanmar,
United Nations Development Programme.
[5] MOECAF (2012), Myanmar’s National Adaptation
Programme of Action (NAPA) to Climate Change, Ministry
of Environmental Conservation and Forestry, Myanmar.
[6] FAO (2017), Myanmar Climate-Smart Agriculture Strategy,
Food and Agriculture Organization of the United Nations.
[Online]. Available:
http://www.fao.org/3/i8465en/I8465EN.pdf
[7] ADB (2019), Myanmar: Climate Change and Water
Resources, Asian Development Bank. [Online]. Available:
https://www.adb.org
[8] Hussain F, Nabi G, Waseem Boota M. (2015). Rainfall
Trend Analysis by Using the Mann-Kendall Test & Sen’S
Slope Estimates: A Case Study of District Chakwal Rain
Gauge, Barani Area, Northern Punjab Province, Pakistan.
Sci Int (Lahore);27(4):3159–65.
[9] Mann, H. B., & Kendall, M. G. (1945). Non-parametric tests
against trend. Econometrica, 13(3), 245–259.
https://doi.org/10.2307/1907187
Figure 8. Sen’s Slope trend map of stations with no trend [10] Sen, P. K. (1968). Estimates of the regression coefficient
in precipitation during winter season. based on Kendall’s tau. Journal of the American Statistical
Association, 63(324), 1379–1389
[11] Yue, S., Pilon, P., Phinney, B., & Cavadias, G. (2002). The
CONCLUTION influence of autocorrelation on the ability to detect trend in
hydrological series. Hydrological Processes, 16(9), 1807–
This study analyzed seasonal precipitation trends 1829. https://doi.org/10.1002/hyp.1095
across ten townships in the Central Dry Zone of Myanmar [12] Yue, S., & Wang, C. Y. (2004). The Mann–Kendall Test
using the Mann-Kendall (MK) test and Sen’s slope Modified by Effective Sample Size to Detect Trend in Serially
estimator. The results revealed distinct spatial and seasonal Correlated Hydrological Series. Water Resources
variations in rainfall patterns. During the summer season, Management, 18(3), 201–218.
[13] Helsel, D. R., & Hirsch, R. M. (2002). Statistical Methods in
most stations exhibited either a slight positive trend or no Water Resources (Vol. 323). U.S. Geological Survey,
significant trend, indicating relatively stable rainfall Techniques of Water-Resources Investigations.
conditions. In contrast, the rainy season showed a general [14] Muia VK, Opere AO, Ndunda E, Amwata DA (2024).
increasing trend in precipitation across the majority of Rainfall and temperature trend analysis using Mann–
townships, with the highest rates observed in Naung-U, Kendall and Sen’s slope estimator test in Makueni County,
Yemathin, and Magway. However, Shwebo and Monywa Kenya. J Mater Environ Sci;15(3):349–67.
presented decreasing trends, suggesting localized declines [15] Sun N, Wigmosta MS, Wada Y, Bierkens MFP, Shehabi A
in monsoonal rainfall. The winter season displayed no (2023). An investigation of the changing patterns of rainfall
in the Indravathi Subbasin utilizing the Mann-Kendall and
significant trend at any station, reflecting a consistent but
Sen’s slope methods. IOP Conf Ser Earth Environ Sci;
low level of rainfall during this period. These findings 1173:012036
highlight the complexity of precipitation dynamics within [16] Mulugeta S, Fedler C, Ayana M (2019). Analysis of long-
the Central Dry Zone and underscore the need for localized term trends of annual and seasonal rainfall in the Awash
climate adaptation strategies. Understanding these seasonal River Basin, Ethiopia. Water (Switz);11(7):1498.
and spatial variations is essential for effective planning in [17] Wilks, D. S. (2011), Statistical Methods in the Atmospheric
agriculture, water resource management, and climate Sciences (3rd ed.), A comprehensive textbook covering
resilience. Continued monitoring and targeted interventions statistical tools used in climate science, including non-
will be critical to address the challenges posed by both parametric tests like Mann-Kendall and methods for
detecting trends and variability.
increasing and decreasing rainfall trends in different parts
[18] Helsel, D. R., & Hirsch, R. M. (2002), Statistical Methods in
of the region. Water Resources, U.S. Geological Survey, Focused on
hydrological data analysis with detailed guidance on non-
REFERENCES parametric methods like Mann-Kendall and Sen’s slope.
[19] WMO (World Meteorological Organization). (2008), Guide
[1] IPCC. (2021). Climate Change 2021: The Physical Science to Climatological Practices (WMO-No. 100), Offers
Basis. Contribution of Working Group I to the Sixth procedures and standards for analyzing climatological data,
Assessment Report of the Intergovernmental Panel on including trend analysis and rainfall variability.
Climate Change. Cambridge University Press. [20] Maidment, D. R. (Ed.). (1993), Handbook of Hydrology,
https://doi.org/10.1017/9781009157896 McGraw-Hill, An authoritative reference covering
 hydrologic processes, data management, and analysis
techniques. Ideal for advanced water resource studies.
[21] Burn, D. H., & Elnur, M. A. H. (2002), Detection of
hydrologic trends and variability. Journal of Hydrology,
255(1–4), 107–122.
[22] Yue, S., & Wang, C. Y. (2004). The Mann–Kendall test
modified by effective sample size to detect trend in serially
correlated hydrological series. Water Resources
Management, 18(3), 201–218.
[23] Kumar, V., Jain, S. K., & Singh, Y. (2010). Analysis of long-
term rainfall trends in India. Hydrological Sciences Journal,
55(4), 484–496.
[24] Agbo EP, Nkajoe U, Edet CO. (2023) Comparison of
Mann–Kendall and Şen’s innovative trend method for
climatic parameters over Nigeria’s climatic zones. Clim
Dyn; 60:3385–401.
[25] Ussalu JLM, Bassrei A (2023). Long memory and trend in
time series of precipitation in Mozambique. Theor Appl
Climatol;154(1–2):643–59.
[26] Sumila TCA, Ferraz SET, Durigon A (2023) . Evaluating
possible changes in air temperature and precipitation patterns
in Mozambique by comparing present and future RegCM4
simulation. Meteorology;2(1):15–36.