The use of a statistical model in this study has revealed strong correlation between climatic conditions and incidences (cases) of dengue fever. This statistical analysis revealed new ways to investigate the relationship between climatic tiers and confirmed disease victims from Islamabad and other major cities in Punjab, including Faisalabad, Lahore, and Rawalpindi.
A total of 18676 cases were reported from Punjab during the year 2019, from January 1st to December 31st. Out of the 8676 reported cases, 31 were reported in Faisalabad, 361 in Lahore, 2893 in Rawalpindi, and 4078 in Islamabad. The highest number of incidences (14) were reported in August 2019 from Faisalabad, 188 in October 2019 from Lahore, 1395 cases in the post-monsoon months from Rawalpindi, and 2571 in the post-monsoon months of 2019 from Islamabad. As shown in (Fig. 1), Islamabad had the highest number of reported cases in 2019 when compared to other cities such as Rawalpindi, Lahore, and Faisalabad.
The highest maximum temperature recorded in Faisalabad during the year 2019 was 40.7°C in June. The lowest minimum temperature was 4°C in January. The maximum rainfall was 87.78 mm in August, and the highest percentage of humidity and number of cases were also recorded in August (78% and 14, respectively). The highest number of reported cases in Lahore were 188, and the highest average monthly rainfall was 400.93mm in October 2019 and the highest average monthly humidity was 78% in December 2019. In Rawalpindi, the highest number of cases (1,395) and the highest average rainfall (400.93mm) were both recorded in October 2019. In Islamabad, the highest number of cases (2571) were reported in October 2019, as well as the highest average rainfall (494.51mm) (Table no. 1).
Figure 1:
Table 1
Table 1
Monthly reported incidences and average climatic factors (temperature, rainfall, and humidity) of Faisalabad, Lahore, Rawalpindi, and Islamabad.
Variables
|
Faisalabad
|
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
Rainfall
|
14.64
|
15.6
|
33.23
|
37.83
|
69.03
|
55.01
|
29.62
|
87.73
|
17.7
|
44.43
|
5.41
|
4.02
|
Temperature
|
Max.
|
19.4
|
22.4
|
27.6
|
33.8
|
38.9
|
40.7
|
37.3
|
36.3
|
36
|
33.6
|
27.4
|
21.8
|
Min.
|
4.4
|
7.4
|
12.6
|
18.1
|
23.3
|
27.4
|
27.3
|
26.9
|
24.2
|
17.6
|
10.4
|
5.7
|
Avg.
|
11.9
|
14.9
|
20.1
|
25.95
|
31.1
|
34.05
|
32.3
|
31.6
|
30.1
|
25.6
|
18.9
|
13.75
|
Humidity
|
60
|
60
|
55
|
47
|
40
|
42
|
74
|
78
|
71
|
67
|
67
|
65
|
Reported cases
|
0
|
0
|
0
|
0
|
5
|
1
|
0
|
14
|
2
|
6
|
2
|
1
|
Variables
|
Lahore
|
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
Rainfall
|
32.63
|
12.85
|
83.75
|
54.53
|
23.1
|
27.63
|
28.54
|
13.71
|
110.56
|
400.93
|
106.15
|
3.82
|
Temperature
|
Max.
|
19.8
|
22
|
27.1
|
33.9
|
38.6
|
40.4
|
36.1
|
35
|
35
|
32.9
|
27.4
|
21.6
|
Min.
|
5.9
|
8.9
|
14
|
19.6
|
23.7
|
27.4
|
26.9
|
26.4
|
24.2
|
18.2
|
11.6
|
6.8
|
Avg.
|
12.85
|
15.45
|
20.55
|
26.75
|
31.15
|
33.9
|
31.5
|
30.7
|
29.6
|
25.55
|
19.5
|
14.2
|
Humidity
|
70
|
71
|
61
|
48
|
39
|
41
|
68
|
71
|
68
|
62
|
65
|
78
|
Reported cases
|
0
|
0
|
4
|
0
|
0
|
2
|
2
|
2
|
39
|
188
|
114
|
10
|
Variables
|
Rawalpindi
|
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
Rainfall
|
23
|
56
|
16
|
44
|
38
|
37
|
65.8
|
100.47
|
237
|
209
|
68
|
36
|
Temperature
|
Max.
|
17
|
19.5
|
24.2
|
29.9
|
35.4
|
39.5
|
35.8
|
33.7
|
33.6
|
30.9
|
25
|
19.3
|
Min.
|
2.7
|
5.5
|
10.4
|
15.3
|
19.9
|
24.5
|
24.8
|
23.6
|
21.6
|
14.5
|
7.5
|
3.3
|
Avg.
|
17
|
19.5
|
24.2
|
29.9
|
35.4
|
39.5
|
35.8
|
33.7
|
33.6
|
30.9
|
25
|
19.3
|
Humidity
|
63
|
57
|
47
|
47
|
49
|
52
|
68
|
72
|
75
|
67
|
59
|
49
|
Reported cases
|
0
|
0
|
0
|
0
|
0
|
2
|
2
|
293
|
1013
|
1395
|
178
|
10
|
Variables
|
Islamabad
|
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
Rainfall
|
136
|
113.44
|
89.8
|
52
|
87
|
31.03
|
42.51
|
97.63
|
252.44
|
494.51
|
91.78
|
16.08
|
Temperature
|
Max.
|
17.7
|
19.1
|
23.9
|
30.1
|
35.3
|
38.7
|
35
|
33.4
|
33.5
|
30.9
|
25.4
|
19.7
|
Min.
|
2.6
|
5.1
|
9.9
|
15
|
19.7
|
23.7
|
24.3
|
23.5
|
26.6
|
13.9
|
7.5
|
3.4
|
Avg.
|
10.15
|
12.1
|
16.9
|
22.5
|
25.57
|
26.8
|
29.65
|
30.6
|
32.89
|
33.43
|
27.9
|
11.55
|
Humidity
|
63
|
57
|
47
|
47
|
49
|
52
|
68
|
72
|
75
|
67
|
59
|
49
|
Reported cases
|
0
|
0
|
0
|
0
|
2
|
0
|
5
|
113
|
1000
|
2571
|
377
|
6
|
Correlation amongst climatic tiers and number of reported incidences
The value of the bivariate correlation analysis indicates how strongly two variables are related. For Rawalpindi, (r = 0.948** P < 0.01)), depicted a strong positive association between monthly average rainfall and number of incidences, (r = 0.613* P < 0.05), depicted a moderate positive association between monthly average humidity and number of incidences, and (r = 0.648* P < 0.05), suggested that there is a moderately positive relationship between monthly average temperature and number of incidences (Table no. 2).
For Lahore in 2019, (r = 0.913** P < 0.01) showed a positive relationship between monthly average rainfall and the number of incidences (Table no. 2).
For Islamabad, (r = 0.961** P < 0.01) showed a strong positive relationship between monthly average rainfall and total number of reported incidences, while (r = 0.683* P < 0.05) showed a moderate positive relationship between monthly average humidity and number of incidences (Table no. 2).
For Faisalabad, there was a moderately positive relationship between monthly average temperature and incidences (r = 0.578* P < 0.05) and a positive relationship between monthly average rainfall and incidences (r = 0.75** P < 0.01) (Table no. 2).
In comparison to other factors like temperature and humidity, average rainfall and the number of reported cases showed the highest correlation in all four cities (FSD, LHR, RWP, and ISL) (r = 0.75**, r = 0.913**, r = 0.948**, and r = 0.961** P < 0.01) respectively. In Rawalpindi and Islamabad (r = 0.948** and r = 0.961** P < 0.01), which had the highest rainfall in comparison to other cities in 2019, more rain resulted in more mosquito breeding and emergence sites, causing an epidemic of this disease in both twin cities (Table no. 2).
Table 2:
Table 2
Temperature, precipitation, and humidity correlation with cases in four major cities (FSD, LHR, RWP, and ISL) for the year 2019.
Variables
|
FSD
|
LHR
|
RWP
|
ISL
|
Rainfall
|
Correlation (r)
|
0.75**
|
0.913**
|
0.948**
|
0.961**
|
% Correlation
|
56.25
|
83.35
|
89.87
|
92.35
|
Average temperature
|
Correlation (r)
|
0.578*
|
0.276
|
0.648*
|
0.531
|
% Correlation
|
33.4
|
7.6
|
41.99
|
28.1
|
Humidity
|
Correlation (r)
|
0.360
|
0.204
|
0.613*
|
0.683*
|
% Correlation
|
12.96
|
4.1
|
37.57
|
46.64
|
Analysis of partial correlation amongst climatic tiers and number of reported incidences
In partial correlation, the strength of an association between two variables is measured while maintaining the effect of one or more variables constant. While keeping the effect of one of the climatic factors, either rainfall, temperature, or humidity, constant for all four cities, the strength of an association between the other two climatic factors was measured using partial correlation.
Effect of average rainfall, while maintaining temperature as the constant variable, on the overall number of reported cases
Zero-order correlations between incidences and average rainfall had values of r = 0.745, r = 0.913, r = 0.948, and r = 0.961; however, when temperature was held constant, the correlation values for FSD, LHR, RWP, and ISL became r12.3=0.576, r12.3=0.914, r12.3=0.913, and r12.3=0.959, respectively. The temperature has little effect on the number of incidences and average rainfall in Lahore, Rawalpindi, and Islamabad, as a negligible difference in values was observed when temperatures were kept constant. Temperature, on the other hand, plays an important role in FSD; by maintaining a constant temperature, the positive correlation between rainfall and the number of incidences decreases. The average temperature in Faisalabad increases Aedes mosquito survival. As a result, the number of incidences decreases while the temperature remains constant (table no. 3).
Effect of average humidity, while maintaining temperature as the constant variable, on the overall number of reported cases
The zero-order correlation between incidences and average humidity was r = 0.360, r = 0.204, r = 0.613, and r = 0.683, respectively; however, when temperature was held constant, the correlation became r12.3=0.41, r12.3=0.403, r12.3=0.357, and r12.3=0.604 for FSD, LHR, RWP, and ISL respectively. Keeping the temperature constant causes a slight increase in the number of incidences for FSD and LHR, while keeping the temperature constant causes a decrease in the number of incidences for RWP, but the temperature has no effect on the relationship between humidity and the number of reported incidences for ISL. In other words, the combination of suitable temperature and humidity acts as a booster for mosquito development and growth, eventually leading to a higher number of incidences (table no. 3).
Effect of average temperature, with humidity as a constant variable, on the total number of reported cases
The zero-order correlation between incidences and average temperature was r = 0.578, r = 0.270, r = 0.648, and r = 0.531, respectively; however, when humidity was held constant, the correlation became r12.3=0.606, r12.3=0.43, r12.3=0.186, and r12.3=0.378 for FSD, LHR, RWP, and ISL. Keeping humidity constant resulted in a slight increase in the positive relationship between average temperature and number of incidences for FSD and LHR, a slight decrease in the association of temperature and incidences for ISL, and a significant decrease in the association of incidences and temperature for RWP, from r = 0.648 to r12.3=0.186. Thus, for RWP, the combination of suitable temperature and humidity acts as a booster for mosquito development and growth, eventually leading to a higher number of incidences (cases) in 2019 (table no. 3).
Effect of average rainfall on the total number of reported cases when humidity is held constant
The zero-order correlation between number of incidences and average rainfall was r = 0.745, r = 0.913, r = 0.948, and r = 0.961, but when humidity was held constant, the correlation became r12.3=0.889, r12.3=0.916, r12.3=0.926, and r12.3=0.926 for FSD, LHR, RWP, and ISL, respectively. The results showed that during the year 2019, humidity alone had no significant effect on the number of incidences and average rainfall, as very few cases were observed during the winter months. In other words, the optimal humidity of 75%, the ideal temperature, and rainfall together act as an amplifier for mosquito growth, development, and transmission, ultimately leading to a higher number of incidents (table no. 3).
Effect of average temperature on the number of incidences when rainfall is held constant
The zero-order correlation between the number of incidences and the average temperature was r = 0.578, r = 0.276, r = 0.648, and r = 0.531, respectively. By controlling for rainfall, the value of correlation became r12.3 = 0.040, r12.3= -0.287, r12.3= -0.19, and r12.3=0.500 for FSD, LHR, RWP, and ISL, respectively. Rainfall has a remarkable effect on the number of incidences and average temperature for FSD; by maintaining constant rainfall, the positive relationship between average temperature and the number of incidences is greatly reduced. For LHR and RWP, the weakly positive relationship between average temperature and the number of incidences changed to a moderately negative relationship when rainfall remained constant. In other words, the lack of precipitation does not create the ideal environment for mosquitoes to survive and grow. As a result, there is a decline in the number of cases (table no. 3).
Effect of average humidity on total number of reported incidences by keeping rainfall constant
The zero-order correlation between the number of incidences and the average humidity was r = 0.360, r = 0.204, r = 0.613, and r = 0.683, but by controlling for rainfall, the correlation became r12.3=0.768, r12.3=0.255, r12.3= -0.380, and r12.3= 0.722 for FSD, LHR, RWP, and ISL, respectively.
Rainfall has a significant impact on the number of incidences and average humidity in Faisalabad, as maintaining constant rainfall increases the positive relationship between average humidity and number of incidences. Keeping rainfall constant had no effect on the association between humidity and total number of reported incidences for LHR and ISL. For RWP, the correlation between humidity and total number of reported cases was r = 0.613; by controlling for rainfall, the correlation became r12.3 = − 0.380. It means that rainfall has a significant impact on the number of incidences and average humidity, as the positive relationship between average humidity and number of incidences decreases when rainfall is constant. As a result, for RWP, the combination of 75% humidity and rainfall acts as an amplifier for mosquito growth, development, and transmission, eventually leading to a higher number of reported incidences (table no. 3).
Table 3:
Table 3
Partial correlation analysis of climatic tiers (temperature, rainfall, humidity) and number of reported incidences.
Constant variables
|
Partial correlation analysis
|
Cities
|
FSD
|
LHR
|
RWP
|
ISL
|
Temperature
|
Zero order (ZO)
|
R
|
0.745
|
0.913
|
0.948
|
0.961
|
Rainfall and incidences (RI)
|
r12.3
|
0.576
|
0.914
|
0.913
|
0.959
|
Zero order (ZO)
|
R
|
0.360
|
0.204
|
0.613
|
0.683
|
Humidity and incidences (HI)
|
r12.3
|
0.417
|
0.403
|
0.357
|
0.604
|
Rainfall
|
Zero order (ZO)
|
R
|
0.578
|
0.276
|
0.648
|
0.531
|
Temperature and incidences (TI)
|
r12.3
|
0.040
|
-0.287
|
-0.19
|
0.500
|
Zero order (ZO)
|
R
|
0.360
|
0.204
|
0.613
|
0.683
|
Humidity and incidences (HI)
|
r12.3
|
0.768
|
0.255
|
-0.380
|
0.722
|
Humidity
|
Zero order (ZO)
|
R
|
0.578
|
0.276
|
0.648
|
0.531
|
Temperature and incidences (TI)
|
r12.3
|
0.606
|
0.439
|
0.186
|
0.378
|
Zero order (ZO)
|
R
|
0.745
|
0.913
|
0.948
|
0.961
|
Rainfall and incidences (RI)
|
r12.3
|
0.889
|
0.916
|
0.926
|
0.926
|