Considering the theme of this research, data analyzed by means of 3 X 3 ANOVA (univariate general linear model), one-way ANOVA, and mean number of infected persons at various neighborhoods classified by the above mentioned criteria of administrative areas and neighborhoods. There is a significant F value in case of neighborhoods, on all months, but not in case of administrative areas (Table 2). As this analysis considered 197 neighborhoods (divided into large cities, medium sized cities, and others including small towns and villages) located at various parts of 13 administrative areas (divided into major, middle sized and other smaller), it could be understood that more than the broader administrative boundaries the locational characteristics matter in case of COVID-19 epidemic. Although, the major administrative areas accounted for the majority of infected cases reported in the country, as a characteristic it plays a role lesser than the more homogenous neighborhoods which is proxy of grass-root level demographics including geography, livelihoods, and proportion of expatriate population, where super-spreaders operate on heterogeneous population, as stated by Eilersen and Sneppen, (2021).
Table 2
Two Way ANOVA results (3X3 Model) calculated with univariate general linear model
Month
|
F (Significance)
|
R2
(Adjusted R2)
|
Corrected
Model
|
Intercept
|
Admin.
area
|
Neighbor-hoods
|
Admin. area X location
|
March
|
6.1
(0.00)
|
11.0
(0.00)
|
0.1
(0.99)
|
15.5
(0.00)
|
0.1
(0.97)
|
0.60
(0.50)
|
April
|
91.3
(0.00)
|
304.1
(0.00)
|
0.7
(0.50)
|
237.7
(0.00)
|
0.8
(0.45)
|
0.86
(0.85)
|
May
|
154.6
(0.00)
|
539.1
(0.00)
|
2.6
(0.07)
|
395.2
(0.00)
|
2.5
(0.08)
|
0.85
(0.85)
|
June
|
111.8
(0.00)
|
427.8
(0.00)
|
4.0
(0.02)
|
272.6
(0.00)
|
3.6
(0.03)
|
0.78
(0.77)
|
July
|
125.5
(0.00)
|
535.6
(0.00)
|
5.4
(0.01)
|
298.5
(0.00)
|
4.8
(0.01)
|
0.79
(0.78)
|
August
|
131.3
(0.00)
|
588.1
(0.00)
|
4.5
(0.01)
|
315.6
(0.00)
|
4.1
(0.02)
|
0.79
(0.79)
|
September
|
140.1
(0.00)
|
633.2
(0.00)
|
4.8
(0.01)
|
336.0
(0.00)
|
4.4
(0.01)
|
0.80
(0.80)
|
October
|
147.3
(0.00)
|
669.4
(0.00)
|
5.0
(0.01)
|
353.2
(0.00)
|
4.6
(0.01)
|
0.81
(0.81)
|
November
|
148.4
(0.00)
|
676.4
(0.00)
|
4.8
(0.01)
|
356.7
(0.00)
|
4.4
(0.01)
|
0.81
(0.81)
|
Moreover, these variables (3X3) are found to have significant interaction in producing the number of COVID- 19 positive cases: similar to the epidemic compartment model tested by Tian et al., (2021). High concentration of population and economic activities in an urban area makes it a hotspot of COVID-19 as revealed by its spread at various locations in the world (Sharifi and Khavarian-Garmsir, 2020). Locations represent climatic conditions, socioeconomic and disease control capacity that determines rate of transmission and health burden (Metelmann et al., 2021). Overall, it implicitly explains that the neighborhood plays a significant role in determining the increase or decrease in number of cases detected on a monthly basis. Statistics are striking that the five major cities having 77.6 percent of infected persons carry only 30.1 percentage of population whereas the middle sized cities and townships having 14.2 per cent of infected persons carry 16.7 percentage of population and other neighborhoods having 8.2 percent of infected persons carry 53.2 percent of the population.
The F value of neighborhood was found to be the highest in May (395.2). Higher F values are reported in November, October, September and August too, which are considered to be months of intense spread of COVID- 19. On the other hand, the F value for administrative area remained low between 0.1 and 5.0 throughout the period: insignificant till May. This pattern was found to be reflected in the interaction also. That means, the two variables operate together to produce differentials on COVID- 19 infections. There are reasonable R2 values, above 0.60 and adjusted R2 above 0.50 making the tests logical.
There are five large cities namely Riyadh, Jeddah, Makkah, Madina, and Dammam, all of them are part of major administrative areas of Riyadh, Makkah Al-Mokarramah, Al-Madina Al-Monawarah, and Eastern Region. These large cities accounted for a large share of the cases but it declined rapidly over the period (Fig. 1). These metropolises with high community and business activities have higher mitigation potential as part of emergency preparedness associated with health delivery; community outreach; diagnostic and surveillance systems etc., as stated by Signorelli, et al., (2020). During the early months of infection, that is, March and April these major cities had higher proportions above 80 percent but which decreased to 48 percent by the month of November. It is to be examined for the effect of intervention strategy adopted, early or late, that impact effectiveness, besides, the characteristics such as spatial architecture and governance pattern (Tian et al., 2021; Ziccardi, 2020). Saudi Arabia has taken a lead in implementing precautionary preventive measures anticipating the danger of epidemic throughout the country (Ministry of health, 2020). Despite, emergencies and disease spread to all parts with varying intensities. Thus, explaining opportunities and dynamics in regional towns are essential, as stated by Guaralda et al., (2020). The future city program initiated in the country is a move on this direction.
On the other hand, there are 21 cities in the country such as Taif (Makkah Al-Mokarramah); Yanbu (Al-Madina Al-Monawarah); Buraiydah (Al-Qassim); AlAhsa, AlJubail, AlMubarass, Dhahran, Hoful. Khobar, Qatif, and Raz Tanura (Eastern Region); Abha, Khammis Mushayt, and Mahayel (Aseer); Tabouk (Tabouk); Hail (Al Hail); Arar (Northern Borders); Jazan (Jazan); Najran (Najran); Baha (Al Baha); and Sakaka (Al-Jouf); whose share has increased. These upcoming neighborhoods (medium sized cities and towns), had multiplied their COVID- 19 cases from 18.0 percent in March to 31.3 percent in November. Here, the impact of major urban components such as environmental quality; socioeconomic impacts; management and governance; and transportation and urban design on COVID-19 spread receives importance (Sharifi and Khavarian-Garmsir, 2020). On the contrary, other 171 smaller neighborhoods also experienced a rapid increase from 2.1 percent in March to 21.0 percent in November. In short, while the larger cities located at major regions had a month wise decline in reported cases, a faster increase in the upcoming and slower increase in smaller neighborhoods experienced. Probably, this trend is reflected in the interaction results.
Separate effects of administrative area and neighborhood were investigated by means of One-Way ANOVA performed on month-wise COVID- 19 cases with administrative areas as well as neighborhoods. The former one is not found to be significant at 0.05 level from May onwards whereas the later variable has significance at 0.00 level throughout the period since March (Table 3). These results are indicative of the above that more than the broader regions, the smaller geographic units play prominent roles in creating the spread of COVID- 19. During the initial stages of infection (March and April), administrative areas did not play significant roles, but slowly their roles became clearer and by September, it started playing the prominence. This explains geographic variations in spread of infection along the population heterogeneity.
Table 3
Results One-Way ANOVA of COVID- 19 month-wise infections with administrative area (3 groups) and neighborhoods (3 groups)
Month
|
Administrative area
|
Neighborhood
|
F
|
Sig
|
F
|
Sig
|
March
|
1.3
|
0.287
|
21.1
|
0.000
|
April
|
2.7
|
0.073
|
279.0
|
0.000
|
May
|
3.2
|
0.044
|
455.5
|
0.000
|
June
|
3.8
|
0.024
|
323.5
|
0.000
|
July
|
4.3
|
0.016
|
357.5
|
0.000
|
August
|
4.0
|
0.020
|
378.8
|
0.000
|
September
|
4.1
|
0.018
|
402.9
|
0.000
|
October
|
4.1
|
0.017
|
422.7
|
0.000
|
November
|
4.1
|
0.017
|
326.9
|
0.000
|
For a better clarity, an analysis of mean number of infected cases was done by taking the model of 3X3 design (Table 4), giving mean for each category of neighborhood (large cities, medium sized cities and townships, and others) in each type of administrative area (major, middle sized, and others). Each neighborhood, as a whole has 1,657 infected persons, as on 28 November, which varied from 33700 in large cities, 5,259 in medium sized cities and townships to 396 in others. In case of large cities, average number of infected persons increased from a low of 176 in March bringing a faster increase at populated industrial cities, the epic centers, as compared to that of mediums sized cities and others. This might be driven by super-spreaders as stated by Eilersen and Sneppen (2021); Tian et al., (2021). This calls for seeking options of human settlements to distribute population including migrants to various parts of the country (Guaralda, et al., 2020).
Table 4
Mean number of infected persons in each category of neighborhood by administrative area classified, month-wise
|
Mar.
31
|
Apr.
30
|
May
31
|
Jun.
30
|
Jul.
31
|
Aug.
31
|
Sep.
30
|
Oct.
31
|
Nov.
28
|
Major administrative areas
|
Large cities
|
176
|
3564
|
12827
|
24168
|
28943
|
30591
|
31913
|
32859
|
33700
|
Medium sized cities
|
15
|
263
|
1214
|
3676
|
5781
|
6248
|
6661
|
6914
|
7044
|
Others
|
1
|
9
|
44
|
157
|
297
|
367
|
396
|
417
|
434
|
Total
|
54
|
418
|
955
|
1844
|
2426
|
2600
|
2739
|
2834
|
2907
|
Middle sized administrative areas
|
Medium sized cities
|
7
|
49
|
216
|
399
|
3116
|
669
|
3925
|
4131
|
4263
|
Others
|
5
|
14
|
28
|
108
|
303
|
425
|
456
|
480
|
498
|
Total
|
6
|
24
|
55
|
254
|
610
|
778
|
834
|
878
|
909
|
Other administrative areas
|
Medium sized cities
|
8
|
71
|
245
|
698
|
1641
|
2385
|
2629
|
2763
|
2885
|
Others
|
1
|
18
|
36
|
62
|
131
|
218
|
239
|
250
|
258
|
Total
|
6
|
28
|
59
|
118
|
253
|
390
|
426
|
446
|
463
|
Total
|
Large cities
|
176
|
3564
|
12827
|
24168
|
28943
|
30591
|
31913
|
32859
|
33700
|
Medium sized cities
|
12
|
151
|
698
|
2316
|
4034
|
4592
|
4919
|
5130
|
5259
|
Others
|
3
|
13
|
38
|
116
|
247
|
335
|
362
|
381
|
395
|
Total
|
36
|
224
|
496
|
944
|
1309
|
1471
|
1552
|
1612
|
1657
|
In case of major administrative areas, the mean number of infected persons, by November, is 2907 with 7044 in medium sized cities and 434 in others: that of large cities remain the same as all 5 of them falls into major administrative areas. The medium sized cities and others of this group of administrative areas had higher number of infected persons than the total. The middle sized administrative areas had only medium sized cities and other neighborhoods, together having 909 mean infected persons (4,263, in medium sized cities and 498 in other neighborhoods). Here, in this category, the medium sized cities and townships undergone rapid increase in the reported cases as compared to the other neighborhoods. In other type of administrative areas, which are relatively smaller, there are fewer mean number of cases, as on November, 463, varying from 2,885 in medium sized cities to 258 in other neighborhoods. Here too, the medium sized cities, reportedly, shows higher rate of increase. Thus, a demand for mitigation strategies at venues where people meet in large numbers of strangers is mandatory, as stated by Eilersen and Sneppen (2021). Besides, different strategies and restrictions adopted shall comply with population density component coupled with employment.
Thus, there arises a need for rethinking beyond conventional growth strategies of cities in line with growth models and urbanization in the emergency preparedness and epidemic spread accounting for planning, design, and development strategies (Guaralda et al., 2020). One way, COVID − 19 enlightened the development community with the planners and policy makers on transformative actions towards creating resilient and sustainable cities (Sharifi and Khavarian-Garmsir, 2020.). It is also important for the authorities to be vigilant and evidence informed as a preparation for immediate disease prevention measures and policies (Metelmann, et al., 2021).