^{*}

The main objective of
this study is to investigate temperature trend and distribution over 29 years
period (1985 to 2013) in Makkah, Saudi Arabia, the holiest city for all Muslims.
Monthly mean, maximum, and minimum temperature levels and their trends are
investigated using Regression analysis and Theil-Sen nonparametric test. Also,
trends in deviations from the reference period (1985-2013) are analyzed. The
results showed that the number of hot days and nights increased annually by
1.5966 and 1.832, respectively, while the number of cold nights decreased
annually by 0.4054 nights. Both Regression analysis and Theil-Sen test
demonstrated positive trends in mean, minimum and maximum temperature levels.
Trends are determined for various seasons and months of the year. The annual
mean of daily mean, maximum and minimum temperature increased by 0.0398^{°}C,
0.0552^{°}C, 0.0398^{°}C per year, respectively. The minimum value of monthly mean
temperature (Tmmean = 23.98^{°}C) was found in January, whereas the maximum value
of the mean temperature (Tmmean= 35.95^{°}C) was found in July. Maximum value of
monthly mean of daily maximum temperature (Tmmax = 43.88^{°}C) was found in June
and minimum (30.54^{°}C) in January. The monthly mean of the daily minimum
temperature (Tmmin) varied between a minimum of 18.82^{°}C in January and a
maximum of 29.59^{°}C in August. From the above analysis it can be concluded that
Makkah is suffering from a considerable warming temperature trend which is
confirmed by the Theil-Sen non-parametric test and there is potentially an
increasing medical risk from heat waves that will be more intense. This
requires specific attention toward: the energy demands for extra cooling, water
resources, draughts, and medical preparedness by the decision makers in order
to minimize these risks to residents, pilgrims who gather annually to perform
hajj rituals and other visitors.

Global warming is the most significant environmental problem the world is experiencing today as concluded by the four IPCC Reports [

Other regional studies over the south Mediterranean [

Another study [

A recent seasonal climate study of temperature in the Arabian Peninsula [

The extreme temperature trends over Jeddah, 70 km away of Makkah, had been analyzed by [

The surface air temperature data of 19 meteorological stations distributed through the KSA using cumulative sum, cumulative annual mean, and the Mann-Kendall rank statistical test for the period of 1978-2010 had showed that there was a negative temperature trend (cooling) with 0.03˚C/year for all stations during the first period (1978-1997), followed by a positive trend (warming) 0.06˚C/year in the second period (1998-2010) with reference to the entire period of analysis [

Kingdom of Saudi Arabia (KSA) occupies about 86% of the area of the Arabian Peninsula (AP) [

spread throughout AP, hence the climate of KSA can be representative for the AP climate [

Makkah (latitude: 21.4 degree North, longitude: 39.85 degree East) has an area of 153,128 km^{2} [

In Makkah the outdoor temperature may exceed 45˚C in summer. Exposes to high outdoor temperatures can result in heat exhaustion or heat stroke to many pilgrims, especially those who are not acclimatized [

Pilgrims in Arafat during the 9th lunar month of Dhu Al Hijjah

Prophet Mohammed travelled to Medina). During that annual event, a drastic increase occurred in the numbers of Pilgrims to Makkah, approximately four millions in 2013. Muslims of more than 80 different nationalities come to Makkah to perform hajj every year [

The extreme hot weather is usually named as heat waves. However there is no universal definition for the heat waves, which is measured relative to the usual weather in the area and relative to the normal temperature for the season. Temperature that people from a hotter climate (acclimatized People) consider normal can be termed a heat wave in a cooler area for (non-Acclimatized). But during the pilgrimage season, most of the visitors belong to different nationalities and some of them who are not acclimatized surly are vulnerable to the temperature increase.

Makkah is located in the south western part of Saudi Arabia (

The effect of urbanization on the air temperature rise has been investigated by [

“Quality Control” is all methods used to detect and remove or reduce errors, loss, incompletion, redundancy, misidentification, misattribution and contamination in the data in the process of recording, manipulating, formatting, transmitting and archiving data to have higher quality, more efficiently and more consistently observation dataset [

Following [

Check of plausibility: to reject those values which never can exist; for example; the negative values of temperature in Makkah; the daily maximum temperature which is less than daily minimum temperature etc.

Calculating the mean climatological value N (norm) of the temperature, and the standard deviation from the norm (s); If the ratio

Gross error limit checks: to determine whether the temperature values fall within a physically realistic range, this ranges is considered to be the mean of the value for the day plus (in case of maximum limit) or minus (in case of minimum limit) four times the standard deviation of the value for the day (s):

Maximum limit = mean + 4(s)

Minimum limit = mean – 4(s)

So the temperature value should fall between the Minimum limit and Maximum limit:

Minimum limit < Temperature Value < Maximum limit.

The daily temperature values outside of these thresholds are marked as potentially problematic; they are manually checked and corrected.

The probability distribution of the minimum, maximum and mean data are assessed using an R-based program and the results shown in

Location of Makkah on Saudi Arabia map

Histogram of the daily minimum temperature (˚C)

Histogram of the daily maximum temperature (˚C)

Histogram of the daily mean temperature (˚C)

(

(

(

The magnitude of the trends of increasing or decreasing temperature were derived from the slopes of the regression line using the least square method and the nonparametric Theil-Sen [

(

From (

The number of hot and cold nights and days were estimated using daily maximum and minimum temperatures recorded during different year. Days are considered hot if the maximum daily temperature exceed 35˚C (Tdmax ³ 35˚C), nights are defined hot when daily min temperature reaches 20˚C (Tdmin ³ 20˚C), the days are defined as cold when Tdmax £ 20˚C and finally nights are classified as cold when Tdmin £ 15˚C [

The data of daily maximum (Tdmax), daily mean (Tdmean) and daily minimum (Tdmin) levels of temperature, monthly mean of maximum (Tmmax), monthly mean of daily mean(Tmmean) and monthly mean of minimum (Tmmin) values of temperatures and annual mean of maximum (Tamax), mean (Tamean) and minimum(Tamin) values of temperatures are analyzed and discussed in the coming sections.

Daily maximum, minimum and mean temperatures during the period (1985-2013) over Makkah

. Linear regression equation for the daily maximum, minimum and mean temperature

Temperature (˚C) | Regression Line | R^{2} |
---|---|---|

Daily maximum (Tdmax) | y = 0.0002x + 31.801 | 0.0105 |

Daily minimum (Tdmin) | y = 0.0003x + 15.736 | 0.0292 |

Daily mean (Tdmean) | y = 0.0001x + 26.085 | 0.0072 |

The overall variations of maximum, mean and minimum levels of daily maximum, daily mean and daily minimum temperature during 1985-2013 are summarized in

The number of hot days (Tdmax ³ 35˚C) and hot nights (Tdmin ³ 20˚C) and cold days (Tdmax £ 20˚C), and cold nights (Tdmin £ 15˚C) during the period of study (1985-2013) were calculated and depicted in (

The long term monthly mean (Tmmean) temperature were calculated using daily average values during the period 1985 to 2013 and were shown in (

. Mean temperature data summary for Makkah during 1985-2013

Label | Max (˚C) | Mean (˚C) | Min (˚C) | Std. Dev. (˚C) |
---|---|---|---|---|

Tdmax | 51.4 | 38.26 | 20.8 | 5.31 |

Tdmean | 45.0 | 30.91 | 15.6 | 4.76 |

Tdmin | 37.0 | 24.84 | 10.6 | 4.48 |

Annual frequency of hot days and nights and cold days and nights

Higher values of mean temperature were observed from May to September, as shown in (

The monthly mean temperature (Tmmean), the corresponding standard deviations, difference between the monthly maximum and minimum temperature of daily mean values (range) and the covariance are given in

It is found that the higher values of covariance (COV) correspond to higher standard deviations (SD) and smaller values of covariance to smaller standard deviations.

Higher values of COV and SD were observed for the winter months and the lower values for summer which is an indicative for relatively more stable temperatures in summer. COV varied between 0.19% and 8.02% corresponding to September and February during the year. This shows that the temperature is most stable in September and least in February.

The trends of monthly mean values of daily mean temperature over different years were obtained using linear regression best fit lines. The linear regression trends for all the months from January to December are shown in (Figures 9(a)-(i)), respectively and the corresponding best fit equations along with coefficient of determination are summarized in

Variations of monthly mean, maximum and minimum of daily mean temperature

. Statistical summary of monthly mean temperature of daily mean values

Month | Mean (˚C) | Std. Dev. (˚C) | Range (˚C) | COV. (%) |
---|---|---|---|---|

Jan. | 23.98 | 1.07 | 4.54 | 1.43 |

Feb. | 24.93 | 1.68 | 7.06 | 8.02 |

Mar. | 27.31 | 1.13 | 4.27 | 2.15 |

Apr. | 30.95 | 0.92 | 3.58 | 2.59 |

May | 34.31 | 0.89 | 3.95 | 2.05 |

Jun. | 35.93 | 0.65 | 2.58 | 3.35 |

Jul. | 35.95 | 0.68 | 3.34 | 2.70 |

Aug. | 35.65 | 0.63 | 2.8 | 2.49 |

Sep. | 34.89 | 0.63 | 2.43 | 0.19 |

Oct. | 32.26 | 0.67 | 3.07 | 3.00 |

Nov. | 28.51 | 0.89 | 3.52 | 2.37 |

Dec. | 25.60 | 1.03 | 3.78 | 3.06 |

. Linear regression equation for all the months (Tmmean)

Month | Regression Line | R^{2} | Month | Regression Line | R^{2} |
---|---|---|---|---|---|

Jan. | y = 0.0205x − 16.924 | R^{2} = 0.0263 | Jul. | y = 0.0385x − 41.041 | R^{2} = 0.2351 |

Feb. | y = 0.1145x − 204.01 | R^{2} = 0.3365 | Aug. | y = 0.0356x − 35.56 | R^{2} = 0.2402 |

Mar. | y = 0.0306x − 33.947 | R^{2} = 0.0535 | Sep. | y = 0.0027x + 29.45 | R^{2} = 0.0013 |

Apr. | y = 0.037x − 43.023 | R^{2} = 0.1179 | Oct. | y = 0.0429x − 53.548 | R^{2} = 0.3017 |

May | y = 0.0292x − 24.123 | R^{2} = 0.0785 | Nov. | y = 0.0339x − 39.271 | R^{2} = 0.1048 |

Jun. | y = 0.0478x − 59.637 | R^{2} = 0.387 | Dec. | y = 0.0437x − 61.753 | R^{2} = 0.1303 |

It is clear from (Figures 9(a) to (c)) that monthly mean of daily mean temperature had increased during the whole months and the annual increases were shown in (

From (

As shown in (

The long term monthly mean (Tmmax) of the daily maximum temperatures (Tdmax) were calculated during the years 1985 to 2013 and shown in (

(

The monthly mean temperature (Tmmax), their corresponding deviations from overall means and standard deviations and COV are given in (

(

Linear regression trends of monthly mean of daily mean temperatures

The annual increase in monthly mean of the daily mean temperature (Tmmean)

Tmmean total increase in the last 29 years

Trend of annual mean of daily mean temperature (Tamean)

Trend of annual deviation from overall mean (Tamean)

Variation of monthly mean, maximum and minimum temperature of daily maximum values

Recorded daily Maximum Temperature which is greater than 44˚C

Linear regression trends of monthly mean of daily maximum temperature

Annual increment in the mean of monthly mean of daily maximum temperature (Tmmax)

Tmmax total increase in the last 29 years

. Statistical summary of monthly mean temperature of daily maximum values

Month | Mean (˚C) | Std. Dev. (˚C) | Mean Dev. (˚C) | Range (˚C) | COV. (%) |
---|---|---|---|---|---|

Jan. | 30.54 | 1.41 | −0.000517241 | 6.54 | 3.46 |

Feb. | 32.05 | 2.02 | −2.75862E-05 | 8.18 | 9.71 |

Mar. | 34.91 | 1.46 | −0.000931034 | 5.95 | 3.37 |

Apr. | 38.69 | 1.09 | −0.000586207 | 3.4 | 4.09 |

May | 42.09 | 1.20 | 3.43021E-15 | 4.75 | 2.87 |

Jun. | 43.89 | 0.80 | −0.000551724 | 3.21 | 2.88 |

Jul. | 43.06 | 0.89 | −4.13793E-05 | 4.05 | 2.94 |

Aug. | 42.77 | 0.80 | −0.000172414 | 3.38 | 1.949 |

Sep. | 42.66 | 0.825 | −3.44828E-05 | 3.22 | 0.61 |

Oct. | 40.18 | 1.06 | −6.89655E-05 | 5.09 | 4.16 |

Nov. | 35.45 | 1.39 | −0.000724138 | 6.22 | 4.20 |

Dec. | 32.15 | 1.49 | −1.10257E-15 | 4.71 | 6.16 |

. Linear regression equation for all the months (Tmmax)

Month | Regression Line | R^{2} | Month | Regression Line | R^{2} |
---|---|---|---|---|---|

Jan. | y = 0.0494x − 68.233 | R^{2} = 0.0896 | Jul. | y = 0.042x − 40.901 | R^{2} = 0.1612 |

Feb. | y = 0.1387x − 245.2 | R^{2} = 0.3417 | Aug. | y = 0.0278x − 12.892 | R^{2} = 0.0875 |

Mar. | y = 0.0481x − 61.32 | R^{2} = 0.0785 | Sep. | y = 0.0087x + 25.33 | R^{2} = 0.008 |

Apr. | y = 0.0585x − 78.239 | R^{2} = 0.2091 | Oct. | y = 0.0594x − 78.635 | R^{2} = 0.2269 |

May. | y = 0.041x − 39.908 | R^{2} = 0.0841 | Nov. | y = 0.06x − 84.39 | R^{2} = 0.1348 |

Jun. | y = 0.0412x − 38.476 | R^{2} = 0.1904 | Dec. | y = 0.0619x − 103.38 | R^{2} = 0.3112 |

As shown in (

(

The monthly mean temperatures along with the monthly maximum and minimum of daily minimum (Tmmin) values during the period 1985 to 2013 are shown in (

The monthly maximum of daily minimum temperatures (37˚C) was recorded on the 2nd of June 2012 while the monthly minimum of daily minimum temperature (10.6˚C) was recorded on the 8th of February 1993 as shown in (

The monthly mean of daily minimum temperature (Tmmin), the corresponding standard deviations from overall mean, the range, and the COV are given in (

The deviations of monthly mean temperature from overall mean show negative values for all months the highest values were observed in January −0.0000586207˚C and lowest in July −0.000931034˚C.

The maximum range of 7.54˚C was observed for February and minimum of 3.45˚C in September.

Trend of annual mean of daily mean temperature (Tamax)

Trend of annual deviation from overall mean (Tamax)

Variation of monthly mean, maximum and minimum temperature of daily minimum values

Variability of daily minimum temperature during the period 1985 to 2013

. Statistical summary of monthly mean temperature of daily minimum values

Month | Mean (˚C) | Std. Dev. (˚C) | Mean Dev. (˚C) | Range (˚C) | COV (%) |
---|---|---|---|---|---|

Jan. | 18.82 | 1.118806432 | −5.86207E-05 | 4.6 | 3.891 |

Feb. | 19.31 | 1.819478895 | −5.51724E-05 | 7.54 | 10.437- |

Mar. | 21.19 | 1.156291741 | −0.000344828 | 5.13 | 4.897 |

Apr. | 24.54 | 1.210178609 | −0.000137931 | 4.76 | 5.416 |

May | 27.71 | 1.162200753 | −0.000655172 | 5.33 | 4.735 |

Jun. | 28.75 | 1.203095351 | −0.000931034 | 4.26 | 8.506 |

Jul. | 29.31 | 1.129394319 | −9.31034E-05 | 4.47 | 7.52 |

Aug. | 29.60 | 1.152320854 | −0.000275862 | 4.71 | 7.347 |

Sep. | 28.82 | 0.824892738 | −0.00062069 | 3.47 | 3.214 |

Oct. | 26.08 | 0.900352695 | −0.000241379 | 3.55 | 5.828 |

Nov. | 23.10 | 0.887764115 | −0.000137931 | 3.62 | 4.667 |

Dec. | 20.38 | 0.944911574 | −0.000103448 | 3.91 | 4.334 |

The linear regression trends of monthly mean of daily minimum temperatures from January to December are shown in (Figures 23(a) to (i)), the corresponding best fit equations in (

The annual mean of daily minimum temperature, as shown in (

Temporal trends of the observed daily minimum, maximum and mean temperature in Makkah have been analyzed for the last 29 years to determine how they are changed over the time using the Theil-Sen nonparametric statistical approach. The advantage of using the Theil-Sen estimator is that it tends to yield accurate confidence intervals even with non-normal data and heteroscedasticity (non-constant error variance). It is also resistant to outliers [

(

(

. Linear regression equation for all the months (Tmmin)

Month | Regression Line | R^{2} | Month | Regression Line | R^{2} |
---|---|---|---|---|---|

Jan. | y = 0.0556x − 92.294 | R^{2} = 0.179 | Jul. | y = 0.1075x − 185.59 | R^{2} = 0.6569 |

Feb. | y = 0.1491x − 278.75 | R^{2} = 0.4869 | Aug. | y = 0.105x − 180.21 | R^{2} = 0.6015 |

Mar. | y = 0.07x − 118.65 | R^{2} = 0.2654 | Sep. | y = 0.0459x − 62.958 | R^{2} = 0.2246 |

Apr. | y = 0.0774x − 130.12 | R^{2} = 0.2963 | Oct. | y = 0.0833x − 140.36 | R^{2} = 0.62 |

May | y = 0.0677x − 107.52 | R^{2} = 0.2456 | Nov. | y = 0.0667x − 110.18 | R^{2} = 0.4089 |

Jun. | y = 0.1215x − 214.17 | R^{2} = 0.7396 | Dec. | y = 0.0619x − 103.38 | R^{2} = 0.3112 |

Linear regression trends of monthly mean of daily minimum temperature

Annual increment in the monthly mean of daily minimum temperature (Tmmin)

Tmmin total increase in the last 29 years

Trend of annual mean of daily minimum temperature (Tamin)

Trend of annual deviation from overall mean (Tamin)

The summer (JJA) season (top left panel) trend is 0.05˚C per year and the 95% confidence intervals in the slope from 0.04˚C/year - 0.06˚C/year. The autumn (SON) season (bottom left panel) trend is 0.05˚C per year and the 95%confidence intervals in the slope from 0.04˚C/year - 0.07˚C/year. The winter (DJF) season (bottom right panel) trend is 0.04˚C per year and the 95% confidence intervals in the slope from 0.02˚C/year - 0.06˚C/year. In all panels, the *** show that the trend is significant to the 0.001 level. The significance level in this case is very high providing very strong evidence that the seasonal mean temperature increased over the period.

Annual trend in daily mean temperature

Seasonal trend in daily mean temperature

(

Monthly trend in daily mean temperature

Annual trend in daily maximum temperature

(

(

(

Seasonal trend in daily maximum temperature

Monthly trend in daily maximum temperature

Annual trend in daily minimum temperature

slope from 0.08˚C/year - 0.1˚C/year. The *** shows that the trend is significant to the 0.001 level. The significance level in this case is very high providing very strong evidence that the annual minimum temperature increased over the period.

(

(

The main findings of this study could be summarized in the following points:

• The number of hot days/nights increased annually by 1.5966/1.832 days and during the period 1985 to 2013 and during the last 29 years the hot days/nights have increased by 45.30/53.128 days and confirmed by [

• The number of cold nights decreased annually by 0.4054 which implies that the number of cold nights were decreased with 11.7566 days during the last 29 years.

• The regression trend analysis of monthly mean temperatures Tmmean, Tmmax and Tmmin showed positive trend in temperature in Makkah and the covariance values indicated more stable temperature patterns in summer.

• The minimum value of the monthly mean of the daily mean temperature (Tmmean) of 23.98˚C was found in January while a maximum value of 35.95˚C was recorded in July.

Seasonal trend in daily minimum temperature

Monthly trend in daily minimum temperature

• The monthly mean of daily mean temperature have increased during the whole months.

• Tmmean increased by 3.3205˚C, 1.3862˚C, 1.2673˚C, 1.2441˚C, 1.1165˚C, and 1.073˚C during the months February, June, December, October, July and April of the whole period.

• The trend of the annual mean of daily mean temperature increased by 0.0398˚C per year, which implies that over the last 29 years the annual mean temperature of Makkah has increased by 1.1542˚C.

• The maximum value of Tmmax of 43.88˚C was found in June while a minimum of 30.54˚C in January.

• The number of days of recoded temperatures exceeds 44˚C increased 0.0002 day annually during the period.

• The monthly mean of daily maximum temperature have increased with 1.4326˚C, 4.0223˚C, 1.3949˚C, 1.6965˚C, 1.189˚C, 1.1948˚C, 1.218˚C, 0.8062˚C, 0.2523˚C, 1.7226˚C, 1.74˚C and 1.7951˚C during the last 29 years

• The annual mean of daily maximum temperature show an increasing trend with an annual rise of 0.0552˚C, which implies that over the last 29 years the annual mean of daily maximum temperature of Makkah has increased by 1.6008˚C.

• The Tmmin varied between a minimum of 18.82˚C in January and a maximum of 29.59˚C in August.

• The monthly maximum of daily minimum temperatures (37˚C) was recorded on the 2nd of June 2012 while the monthly minimum of daily minimum temperature (10.6˚C) was recorded on the 8th of February 1993.

• The Tmmin has increased during the last 27 years with 1.6124˚C, 4.3239˚C, 2.03˚C, 2.2446˚C, 1.9633˚C, 3.5235˚C, 3.1175˚C, 3.045˚C, 1.3311˚C, 2.4157˚C, 1.9343˚C and 1.7951˚C.

• The increasing trend in the annual and the monthly mean of daily mean (Tdmean), daily maximum (Tdmax) and daily minimum (Tdmin) temperatures (Tmmin) determined by regression method have been confirmed by the nonparametric Theil-Sen method.

• The increasing trend of the annual mean of minimum, maximum and mean temperature is confirmed by both [

The analysis of the hot days/nights based on temperature thresholds reveals that summers are expanding and winters are shrinking in Makkah, resulting in more pressure on water and energy sectors. The increasing trend of intense heat may cause health problems for the pilgrims. The observed trend of temperature increase is alarming; it may alter the local climate as observed very recently in terms of flash flooding in Makkah. The results indicate the vulnerability of the Holly City. The results will be helpful for the policy makers to reduce the future risks associated with rapidly changing climate of Makkah.

Thanks are expressed to the Presidency of Meteorology and Environment in Saudi Arabia for providing the observation dataset. Also, the author would like to acknowledge Dr. Bakry Ben Maatouq Assas (the Director of Um-AlQura University), Prof. Dr. A. H. Asghar (the Dean of the Custodian of the Two Holy Mosques’ Institutes for Haj and Umrah Research), Dr. Turki Habeebullah (the Head of the Environment and Health Researches Department) for their support and continuous encouragement and Dr. Said Munir for internally reviewing the manuscript. The author greatly appreciates the contribution of the anonymous reviewers, as a result of which the manuscript has considerably improved.