Sample Paper - Manuscript Preparation


16 
J. mt. area res., Vol. 4, 2019 

   

       Journal of Mountain Area Research 
            

 

 

ANALYSIS OF PRECIPITATION TRENDS IN GILGIT-BALTISTAN (GB), 

PAKISTAN 
M. Raza* 

Department of Physics, Karakorum International University Gilgit-Baltistan 

 

ABSTRACT 

This article explores seasonal as well as annual trends of precipitation in Gilgit-Baltistan (GB) Pakistan 

over the period between 1980 and 2012. For the analysis of precipitation trends, non-parametric 

Mann-Kendall and Sen’s tests were applied. The assessment of precipitation data showed an annual 

increasing trend of precipitation in GB although this trend is not statistically significant. Seasonal 

analysis revealed least amount of precipitation occurs in the autumn season which is becoming drier. 

Trend analysis showed an increasing trend of precipitation in the other three seasons (winter, Spring 

and summer). The study revealed that the precipitation at Gupis is increasing significantly in winter, 

spring and summer seasons. 

KEYWORDS: Precipitation, Mann-Kendall, significant, Gilgit-Baltistan. 

 

*Corresponding author: (Email: raza.physics@kiu.edu.pk) 

1. INTRODUCTION 

Winter and summer are the two major 

rainy/precipitation seasons in Pakistan. In winter, 

the mid-latitude westerly waves move across 

the lower latitudes and their troughs generally 

extend down to 35N, and sometimes even 

further south. Under the influence of westerly 

waves as well frontal system, the northern parts 

of Pakistan receive substantial rainfall over low 

elevation and snow fall at high elevation areas 

in the winter season. In Pakistan the total annual 

precipitation ranges from 500 mm to 800 mm. 

The northern half of the country receives 

handsome amount of precipitation in both 

winter and summer, but the southern half 

receives hardly 50% of the amount received in 

the north [1].  

Solid precipitation over the northern 

mountains melts in early summer and maintains 

sustainable river flow for power generation and 

irrigation before the onset of summer monsoon. 

In addition to this precipitation, winter rain 

bearing systems yield substantial rainfall in sub-

mountainous and low elevation plains [2].  

The northern part especially GB is host to the 

world’s largest glaciers and is the major water 

resource the River Indus, the leading river of the 

country [3]. Changing trends of precipitation 

and extreme events in GB bring floods in the 

Indus on large scale and damage the 

population living down the country. In context 

of climate change and disaster management, 

it is therefore important to understand the 

precipitation trends. The aim of this study is to 

analyze the recent trends of annual and 

seasonal precipitation in GB.      

2. DATA AND STUDY AREA 

To carry out precipitation assessment, daily 

precipitation data over the period 1980 - 2012 

from the six weather stations situated in GB were 

obtained from Pakistan Meteorological 

Department (PMD). Fig. 1 and Table 1 show the 

Full length article 

Vol. 4, 2019 

http://journal.kiu.edu.pk/index.php/JMAR 



 

 

Raza et al., J. mt. area res. 04 (2019) 16-23 

17 

J. mt. area res., Vol. 4, 2019 

specific locations of the six stations and data 

period of precipitation used in this study. Using 

this dataset monthly, seasonal and annual total 

precipitation was calculated, and trend analysis 

was also performed.  

For all the six stations homogeneity analysis of 

the dataset was performed on monthly total 

precipitation. Graphical analysis was carried 

out to check the homogeneity of the dataset. 

The monthly totals of all the six stations have 

been used to calculate the regional average 

and the regional precipitation series.  

3. METHODOLOGY 

Monthly total precipitation was calculated 

from the daily precipitation data. These 

monthly totals were used to calculate the 

annual and seasonal precipitation. The 

regional precipitation index was developed 

using monthly precipitation of the six weather 

stations in GB. The stepwise methodology of 

the study is given in Fig.2. 

Table 1: Details of Stations 

General 

Location 

Station 

Name  

Latitude Longitude Average 

Altitude    

a.s.l. 

(meters) 

GB  

Pakistan 

Astor 35o20/ 74o 54/ 2168 

Bunji 35o40/ 74o 38/ 1372 

Chilas 35o25/ 74o 06/ 1251 

Gilgit 35o55/ 74o 20/ 1460 

Gupis 36o10/ 73o 24/ 2155 

Skardu 35o18/ 75o 41/ 2210 

3.1 Mann-Kendall test 

The Mann-Kendall test, which is a non-

parametric; therefore, data outliers do not 

affect the results [4, 5, 6, 7]. It was applied to the 

long-term data in this study to detect statistically 

significant trends. In this test, the null hypothesis 

(H0) was that there has been no trend in 

precipitation over time; the alternate hypothesis 

(H1) was that there has been a trend 

(increasing/decreasing) over time. For the 

assessment of changes in magnitude of 

precipitation we used Sen’s test method. The 

mathematical equations for calculating Mann-

Kendall statistics S, V(S) and standardized test 

statistics Z are follows: 

𝑆 = ∑ ∑ 𝑆𝑔𝑛(𝑥𝑖 − 𝑥𝑗)

𝑛

𝑗=𝑖+1

𝑛−1

𝑖=1

 

(1) 

𝑆𝑔𝑛(𝑥𝑖 − 𝑥𝑗) = {

1       𝑖𝑓 𝑥𝑖 − 𝑥𝑗 > 0

0      𝑖𝑓 𝑥𝑖 − 𝑥𝑗 = 0

−1     𝑖𝑓 𝑥𝑖 − 𝑥𝑗 < 0

 

 (2) 

𝑉(𝑆) =
1

18
[𝑛(𝑛 − 1)(2𝑛 + 5) − ∑𝑡𝑝(𝑡𝑝 − 1)

𝑞

𝑝=1

(2𝑡𝑝

− 5)] 

         (3) 

𝑍 =

{
 
 

 
 

𝑆−1

√𝑉(𝑆)
       𝑖𝑓 𝑆 > 0

0                 𝑖𝑓 𝑆 = 0
𝑆+1

√𝑉(𝑆)
     𝑖𝑓 𝑆 < 0

 

  (4) 

Where xi and xj are the annual values in year 

i and j, i > j, respectively, n is the length of time 

series, tp is the number of ties for pth value, and 

q is the number of tied values. Positive Z values 

indicate an upward trend and negative Z 

values indicate downward trend. If│Z│ > 𝑍1−𝛼/2 

, (H0) is rejected and statistically significant trend 

exist in hydrologic time series. The critical value 



 

 

Raza et al., J. mt. area res. 04 (2019) 16-23 

18  

J. mt. area res., Vol. 4, 2019 

of 𝑍1−𝛼/2  for p value of 0.05 from the standard 

normal table is 1.96. 

 

  Fig. 1:  Study Area and Location of the Meteorological Stations  

3.2 Sen’s test 

To estimate the true slope of an existing trend 

(change per year) the Sen’s nonparametric test 

is used [8]. In this study this method was used to 

estimate the magnitude of trend in the time 

series data: 

𝑇 =
𝑥𝑗−𝑥𝑘

𝑗−𝑘
                    (5) 

In this equation, 𝑥𝑗 and 𝑥𝑘  represent data 

values at time j and k, respectively. Consider 

𝑄𝑖 = {

𝑇(𝑁+1)/2              𝑁 𝑖𝑠 𝑜𝑑𝑑

1

2
(𝑇𝑁

2

+ 𝑇𝑁+2
2

) 𝑁 𝑖𝑠 𝑒𝑣𝑒𝑛
            

(6) 

A positive 𝑄𝑖 value represents an increasing 

trend; a negative 𝑄𝑖  value represents a 

decreasing trend over time. 

 

 

 

 

Fig. 2: The Stepwise Data Organization and the 

Analytical Tools Used in this Study. 

For detecting and estimating trends in the 

time series Mann-Kendall and Sen’s tests were 

performed, using Microsoft Excel template 

MAKESENS, which was developed by the Finnish 

Meteorological Department. MAKESENS is a 

widely used software for detecting and 

researching precipitation trends [4, 5]. For this 

study we divided the seasons of GB into four 



 

 

Raza et al., J. mt. area res. 04 (2019) 16-23 

19 

J. mt. area res., Vol. 4, 2019 

categories. Table 2 describes the seasons and 

respective months.  

Table 2: Seasons of Gilgit-Baltistan 

 

 

 

4. RESULTS 

4.1  Station Basis 

4.1.1 Increasing Trend of Annual 

Precipitation 

Table 3 shows that in the region of GB the 

annual mean precipitation varies by location. 

During the period from1980 to 2012 the annual 

mean precipitation varies from 148 to 485 mm. 

Highest precipitation was recorded in Astor and 

the lowest in Gilgit.  

Fig. 3: Seasonal Trend of Precipitation in GB 

 

Mann-Kendall test results revealed upward 

trend in annual precipitation at five weather 

stations i.e. Bunji, Chilas, Gilgit, Gupis and 

Skardu, and a downward trend of annual 

precipitation at Astor. Annual precipitation is 

increasing significantly at Gupis with a rate of 

7.348 mm/year (Table 4).  

4.1.2 Seasonal Trends 

Seasonal trend of mean precipitation in GB 

for 33 years (1980-2012) is summarizes in Figure 3. 

The figure shows that the highest amount of 

precipitation occurs during winter and lowest 

during autumn seasons. During winter the mean 

precipitation varies from 39 to 203 mm and in 

spring season it varies between 49 to 142 mm 

(Table 5). Being a winter dominant region, the 

highest amount of precipitation occurred 

during the winter season. As compare to the 

other parts of Pakistan, GB receives less 

precipitation in the summer season. During 

summer the amount of precipitation varies 

between 35 to 75mm (Table 5). 

 

 

 

Seasons Months 

Winter 

Season 

November, December, January, 

February,  

Spring March April, May 

Summer June, July, August, September 

Autumn October, November 

0

50

100

150

200

250

P
re

c
ip

it
a
ti

o
n

 i
n

 m
m

Summer                             Winter                        Spring                      Autumn         

Astor Bunji Chilas Gilgit Gupis Skardu



 

 

Raza et al., J. mt. area res. 04 (2019) 16-23 

20  

J. mt. area res., Vol. 4, 2019 

Table 3: Descriptive Statistics of the Annual 

Precipitation in GB for the period 1980-2012 

Table 4: Mann-Kendall and Sen’s Test Results of 

Annual Precipitation Trend for the Six Stations 

Stations MK Test 

(Test Z) 

Trend Result Sen’s Slope 

Estimate Change in 

Precipitation 

(mm/y) 

Astor -1.13  NS -3.055 

Bunji 0.51  NS 0.716 

Chilas 0.85  NS 0.756 

Gilgit 1.13  NS 1.309 

Gupis 3.46  Sig. 7.348 

Skardu 0.95  NS 1.756 

mm= Millimeter;  MK:  Mann-Kendall;  NS:  Not  

Significant;  Sig.:  Significant; ↑: Upward; ↓:Downward 

There is an upward trend of precipitation during 

summer season in the whole of GB region which 

clearly shows the approach of monsoon rain in 

this region. Spring rainfall is also increasing in 

Gilgit, Gupis and Skardu and it is decreasing in 

Astor, Bunji and Chilas. Winter precipitation is 

increasing at all the stations except Astor and 

Chilas where it is decreasing non-significantly. In 

Gupis the winter precipitation is significantly 

increasing with an increasing rate of 2.15 

mm/year. 

Table 5: Seasonal Mean, Standard Deviation and 

Range of Precipitation at Various Stations 

Stations Seasons 

 

 

Mean  

Precip

itation 

(mm) 

Std. 

Dev. 

Precipi

tation 

(mm) 

Range 

Precipitation 

(mm) 

Min Max 

Astor Winter 202.57 83.16 43.3 381.5 

Spring 142.33 84.61 26.4 344.9 

Summer 74.83 40.57 18.1 201.6 

Autumn 39.09 41.85 0.6 170.6 

Bunji Winter 46.24 31.19 5.7 126.7 

Spring 51.90 40.43 6.5 144.3 

Summer 55.63 33.46 0.0 138.0 

Autumn 10.66 22.79 0.0 124.0 

Chilas Winter 82.97 42.88 16.8 180.3 

Spring 66.38 14.24 0.1 263.5 

Summer 41.68 31.80 3.0 152.0 

Autumn 14.24 28.13 0.0 146.2 

Gilgit Winter 39.82 21.44 7.5 108.1 

Spring 49.36 34.70 4.1 116.9 

Summer 40.66 25.15 11.5 121.0 

Autumn 9.40 18.34 0.0 102.4 

Gupis Winter 60.28 53.55 0.0 222.3 

Spring 73.59 95.11 0.0 410.9 

Summer 60.19 50.17 2.3 182.4 

Autumn 12.34 21.74 0.0 109.7 

Skardu Winter  128.32 68.37 39.2 263.0 

Spring  59.73 55.67 3.0 218.0 

Stations Mean 

Precipitation 

(mm) 

Std.Dev.  Range Precipitation 

(mm) 

Min Max 

Astor 484.83 137.51 248.4 887.0 

Bunji 167.12 68.18 63.2 339.8 

Chilas 182.57 99.70 55.1 573.4 

Gilgit 148.23 56.62 67.5 300.9 

Gupis 212.15 169.38 63.0 683.0 

Skardu 267.15 213.68 87.0 1334.0 



 

 

Raza et al., J. mt. area res. 04 (2019) 16-23 

21 

J. mt. area res., Vol. 4, 2019 

Summer  35.46 20.63 0.6 94.0 

Autumn  14.29 25.24 0.0 114.5 

 

Table 6: Seasonal Precipitation in Gilgit-Baltistan 

Region Seasons 

 

 

Mean 

Precipitatio

n (mm) 

Std. 

Dev. 

 

Range 

Precipitation 

(mm) 

Min Max 

GB Winter 94 63 19 214 

Spring 74 35 7 250 

Summer 52 15 6 148 

Autumn 17 11 0.1 128 

Table 7: Trend of Seasonal Precipitation of all Six 

Stations 

Station

s 

Seasons 

 

 

MK Test 

(Test Z) 

 

Trend 

 

Resu

lt 

 

Sen’s Slope 

Estimate 

Change in 

Precipitatio

n (mm/y) 

Astor Winter -0.08  NS -0.15 

Spring -0.33  NS -0.39 

Summer 0.88  NS 0.53 

Autumn -1.27  NS -0.57 

Bunji Winter 0.82  NS 0.42 

Spring -0.50  NS -0.18 

Summer 0.93  NS 0.60 

Autumn 0.30  NS 0.00 

Chilas Winter -1.12  NS -1.03 

Spring -0.64  NS -0.33 

Summer 1.07  NS 0.41 

Autumn -0.08  NS 0.00 

Gilgit Winter 0.62  NS 0.32 

Spring 0.08  NS 0.12 

Summer 0.91  NS 0.12 

Autumn -0.45  NS -0.04 

Gupis Winter 3.56  Sig. 2.15 

Spring 1.80  Sig. 1.22 

Summer 2.18  Sig. 1.49 

Autumn 1.46  NS 0.17 

Skardu Winter 0.76  NS 0.93 

Spring 0.81  NS 0.48 

Summer 1.89  Sig. 0.59 

Autumn -1.57  NS -0.16 

4.2  Regional Trend Analysis 

Table 6 shows the seasonal means, respective 

standard deviations and the range 

precipitation in GB between the period 1980 

and 2012. Seasonal percentages of mean 

precipitation revealed that forty percent of total 

annual precipitation occurs during winter 

season and 31% is received in the pre-monsoon 

season. Maximum amount of precipitation was 

observed during the winter and pre-monsoon. 

Table 7 summarises the results of seasonal trend 

analysis of all stations. The Mann-Kendall test 

shows an upward trend of precipitation during 

the winter and summer seasons and downward 

trend in the autumn season. However the results 

are not statistically significant. Autumn season 

receives the lowest amount of precipitation 

showing the downward trend of precipitation 

which means that the region is becoming drier. 

On the basis of this trend analysis we can predict 

that GB will receive more precipitation in future 

during winter and summer seasons. 

5. DISCUSSIONS 

Upstream water supply is crucial to sustain 

upstream water reservoir system, which are used 

to store and release water to downstream area 

when most needed. Indus Basin Irrigation System 



 

 

Raza et al., J. mt. area res. 04 (2019) 16-23 

22  

J. mt. area res., Vol. 4, 2019 

is the world largest irrigation network, which 

regulated through two major storage dams 

(Tarbela dam on Indus River and Mangla dam 

on the Jhelum River). Both are located in the 

upper Indus basin and are fed predominantly by 

melt water. Any change in upstream water 

supply to these dame will have profound effect 

on millions of people downstream [9]. This study 

investigated variability in monthly, seasonal and 

annual precipitation at 6 stations in the upper 

Indus basin over a 33- year study period (1980-

2012). The mean annual precipitation at 

different stations showed considerable 

variation. Astor showed negative trend 

(decreased precipitation over time) at 5% 

significance level in the annual precipitation 

series. This will affect the contribution of water 

flow in the Indus from Astor basin. At the same 

significance level (5%), at the remaining stations, 

there was an increasing trend in the annual 

precipitation trend. The significant changes in 

seasonal and annual precipitation over time 

was detected at Gupis station. 

These study results follow the same statistical 

trend by Salma Khalid et al. [10] Hartmann et al. 

[11] for Indus River basin. Findings of this study 

were consistent with results from Imran [12], Stein 

Bauer [13], Rasul [14], Dimri [15] wherein they 

found statistically significant increasing trends in 

winter precipitation for northern area of 

Pakistan.  These deviations in precipitation 

trends may bring more water related disasters 

such as drought and flood in the near future. 

CONCLUSIONS 

Following conclusion can be deduced from the 

study: 

seasons; at 214 mm and 250 mm respectively. 

Twenty two percent of the total mean 

precipitation occurs during summer season. The 

lowest precipitation occurs during the autumn 

season i.e. 7% of the total annual precipitation. 

 Total annual precipitation in GB has increased. 

However, the trend is not statistically 

significant. 

 In GB, 40% of the total annual precipitation 

was received in the winter season. The spring 

and summer season received 31% and 22% of 

the total precipitation respectively. Autumn 

season received the lowest amount of 

precipitation i.e. 7% of the annual 

precipitation. 

 Autumn precipitation trend was decreasing in 

almost all the areas of GB and this season is 

getting drier. 

 GB is likely to receive more precipitation in the 

winter and summer seasons. 

ACKNOWLEDGEMENT 

I would like to express our sincere gratitude and 

appreciation to the Pakistan Meteorological 

Department (PMD) for providing data and 

valuable information. 

REFERENCES 

[1]  G. Rasul, A. Mahmood, A. Sadiq and S.I. 

Khan, “Vulnerability of the Indus delta to 

climate change in Pakistan”, Pakistan Journal 

of Meteorology, vol. 8, issue. 16, pp.89-107, 

2012. 

[2]  G. Naheed and D.H. Kazmi, “Seasonal 

variation of rainy days in Pakistan”, Pakistan 

Journal of Meteorology, vol. 9, issue. 18, pp. 

9-13, 2013. 

[3]  M. Raza, D. Hussain, G. Rasul, M. Akbar and 

G. Raza, “Variation of surface temperature 

and precipitation in Gilgit-Baltistan (GB), 

Pakistan from 1955 to 2010”, Journal of 

Biodiversity and Environmental Sciences, vol. 

6, no. 2, pp. 67-73, 2015.  



 

 

Raza et al., J. mt. area res. 04 (2019) 16-23 

23 

J. mt. area res., Vol. 4, 2019 

[4]  Ahmad, D. Tang, T.F. Wang, M. Wang and B. 

Wagan, “Precipitation trends over time using 

Mann-Kendall and Spearman’s rho test in 

Swat river basin, Pakistan”, Advances in 

Meteorology, vol. 2015, 2015. 

[5]  Finnish Meteorological Institutes, MAKESENS-

Application for trend calculation [Internet], 

Finnish Meteorological Institutes. Available 

from http://en.ilmatieteenlaitos.fi/makesens [ 

Accessed 1st July 2015]  

[6]  H.B. Mann, “Nonparametric test against 

trend”, Econometrica, vol. 13, pp.245-259, 

1945. 

[7]  M.D. Rehman and M. Begum, “Application 

of non-parametric test for trend detection of 

rainfall in the largest Island of Bangladesh”, 

ARPN Journal of Earth Sciences, vol. 2, pp. 40-

44, 2013. 

[8]  P.K. Sen, “Estimate of the regression 

coefficient based on Kendall’s tau”, Journal 

of the American Statistical Association, vol. 

63, no. 324, pp. 1379-1389, 1968. 

[9]  W.W. Immerzeel, “Climate change will 

affect the Asian water towers”, Science, vol. 

328, no. 1382, 2010.  

[10] S. Khalid, M. Qasim, D. Farhan, “Hydro-

meteorological Characteristics of Indus River 

Basin at Extreme North of Pakistan. J Earth Sci 

Clim Change, vol. 5, no. 170, 2013. 

[11] H. Hartmann and L. Andresky, “Flooding in 

the Indus river basin- A spatiotemporal 

analysis of precipitation records”, Global and 

Planetary Change, vol. 107, pp. 25-35, 2013. 

[12] A. Imran, Q. Zaman and M. Afzal, “Temporal 

trends in the peak monsoonal precipitation 

events over Northeast Pakistan”, Pakistan 

Journal of Meteorology, vol. 10, no. 19, pp. 

19-30, 2013.     

[13] M.J. Steinbauer and J. Zeidler, “Climate 

change in the northern areas of Pakistan 

impact on glaciers, ecology and livelihood”, 

WWF for Nature-Pakistan Gilgit Conservation 

and Information Center, 2008. 

[14] G. Rasul, “Climate data and modelling 

analysis of the Indus region”, Pakistan 

Meteorological Department (PMD), 2012, 

http://wwfpak.org/ccap/pdf/Climate Data 

Modelling%20An- 

alysis%20of%20the%20Indus%20Ecoregion.pd

f. 

[15] A.P. Dimri, “Surface and upper air fields 

during extreme winter precipitation over the 

western Himalayas”, Pure and Applied 

Geophysics, vol. 163, no. 8, pp. 1679–1698, 

2006. 

 

 

 

 

 This work is licensed under a Creative Commons Attribution 4.0 International License. 

Received: 24 April 2019. Revised/Accepted: 4 June 2019. 

http://en.ilmatieteenlaitos.fi/makesens
http://creativecommons.org/licenses/by/4.0/
http://creativecommons.org/licenses/by/4.0/
http://creativecommons.org/licenses/by/4.0/