Correlation between ‘+ONI’ and ‘rainfall index’
The relationship between + summer ONI (April, May, June) and rainfall index of monsoon (July, August, September) and post monsoon (October, November, December) rainfall is shown in fig 1. During +summer ONI, when there was positive El Niño (+ONI) influence, negative correlation with rainfall (deficit rainfall than normal ) was observed in the states of Haryana, Delhi, Chandigarh, Kerala, West Bengal, Tamilnadu, Goa, Maharashtra, coastal Karnataka, East Uttar Pradesh, and Gujarat. Whereas, in the states like Andaman & Nicobar Island, Coastal Andhra Pradesh, East Uttar Pradesh, North Interior Karnataka, Jammu and Kashmir a positive correlation was found indicating rainfall in excess of normal. A negligible relationship was found in the states of Arunachal Pradesh, Maharashtra, east Rajasthan, east and west Madhya Pradesh.
The association between + monsoon ONI ( July, August, September) and ‘rainfall index (October, November, December), negative correlation was found in the states of Assam, Nagaland, Andaman & Nicobar Islands, Meghalaya, West Bengal, Tripura, Manipur, South Interior Karnataka, Tamilnadu, Gujarat, and Pondicherry indicating rainfall deficit than normal. Whereas, the states like Haryana, Delhi, Chandigarh, Himachal Pradesh, Jammu and Kashmir, Punjab, North Interior Karnataka, Uttarakhand, and Vidarbha region showed positive correlation indicating excess rainfall than normal. Insignificant relationship was found in Madhya Maharashtra, Lakshadweep, Chhattisgarh, Bihar, Telangana, and West Rajasthan.
Correlation between ‘rainfall index’ and ‘dengue case index’
The correlation between rainfall index (monsoon season; July, August, September) and ‘dengue case index’(July, August, September) show negative correlation that most of the states like Haryana, Gujarat, Himachal Pradesh, Chhattisgarh, Chandigarh and Delhi, A & N Islands, Konkan and Goa, Madhya Pradesh, Maharashtra, Nagaland, Manipur, Mizoram, Tripura, Tamil Nadu and Pondicherry, Uttar Pradesh, Uttarakhand, Chattishgarh, and West Bengal register negative correlation (fig. 2), indicating an increase in dengue with rainfall deficit in the monsoon season. However, when the analyses in respect of Delhi was done with lag periods, it is found that 1-3 month lag between ‘rainfall index’ and ‘dengue case index’. On the other hand in the states like Arunachal Pradesh, Bihar, Jharkhand, Karnataka, Rajasthan, Andhra Pradesh & Orissa, a positive correlation was found indicating dengue outbreaks during excess rainfall. No correlation was found in the states like Kerala, Punjab, Assam, and Meghalaya.
The mapping of correlation coefficient between + monsoon ONI (July, August, and September) and ‘Dengue case index’ (October, November, and December) is shown in fig. 3b. A positive correlation (0.40-0.68) was found in the states of Arunachal Pradesh, Chhattisgarh, Haryana, A & N Islands, Daman & Diu, Delhi & Haryana indicating that with + monsoon ONI increase in ‘dengue cases’ is likely to be witnessed in post monsoon months i.e. October, November and December. Whereas negative correlation was found in the states of Assam, Himachal Pradesh, Andhra Pradesh, Jharkhand, Meghalaya, Manipur, Nagaland, Sikkim and Mizoram reflecting decrease in dengue cases after + monsoon ONI (monsoon El Niño). Rest of the states show ‘very low or no correlation’.
Predicted Dengue outbreak states based on ENSO outlook of 2019
As per probabilistic ENSO outlook (CPC/IRI) (fig. 4) extracted in January 2019, there are 55-72% chances of moderate El Niño during April to June (summer season) indicating the possibility of moderate rainfall deficit during July to September. Therefore, the states of Haryana, Himachal Pradesh, Assam, Gujarat, Rajasthan, Delhi, Mizoram, Telangana, Madhya Pradesh, Jharkhand, Uttar Pradesh, Tripura, and Manipur are likely to have dengue outbreaks during July to September month of 2019. (fig. 5a).
From July onwards the chances of low intensity of El-Niño (less than 50%) indicates the probability of below-normal rainfall from July to September. The states of Rajasthan, Telangana, Haryana, Jharkhand, Andhra Pradesh, and Gujarat are likely to have dengue outbreaks during October to December month of 2019, provided surveillance and interventions remain unchanged (fig. 5b). From June onwards, emerging La Nina conditions are visible, indicating that the states of Tamil Nadu, Maharashtra, Himachal Pradesh, Chhattisgarh and Orissa are likely to have dengue outbreaks as in these states there is negative correlation between +ONI and dengue cases.
Validation
As per results some states (i.e., Arunachal Pradesh, Chhattisgarh, Haryana, Uttarakhand, Andaman and Nicobar Islands, Delhi, Daman and Diu) are positively correlated with ‘+ONI’ while some are negatively correlated (i.e., Assam, Himachal Pradesh, Meghalaya, Manipur, Mizoram, Uttar Pradesh, Orissa, Jammu and Kashmir and Andhra Pradesh). For validation of findings, long term data of one district each from positively and negatively correlated states was showed for the detailed validation of the association between +ONI and dengue cases for a micro-scale study. The results show that the association between +ONI and occurrence of dengue cases in showed districts is statistically significant based on hypothesis test. It indicates that in Delhi and Visakhapatnam, 1-3 months lag between +ONI and reported dengue cases was observed. The results also showed that 2 month lag was found between monthly rainfall index and dengue cases in both the places (Delhi and Visakhapatnam). Lag month-wise correlation values between dengue cases and +ONI, rainfall were shown in table 1. The largest El Niño events (ONI) resulted in very limited no. of dengue cases reported in Visakhapatnam in 2015 (fig. 6) while Delhi recorded a very high no. of dengue cases in 2006, 2009, 2015 (fig. 7).
Table 1. Correlation Coefficients between monthly +ONI and Dengue Cases; and rainfall and Dengue Cases (2014-2018) with different lag periods in Delhi and Visakhapatnam
Lag-months
|
Delhi (Correlation coefficients, ‘r’)
|
Visakhapatnam (Correlation coefficients, ‘r’)
|
+ONI & dengue cases
|
Rainfall & dengue cases
|
+ONI & dengue cases
|
Rainfall & dengue cases
|
0
|
0.368
|
-0.053
|
-0.541
|
0.316
|
1
|
0.429
|
0.189
|
-0.545
|
0.661
|
2
|
0.401
|
0.412
|
-0.552
|
0.742
|
3
|
0.217
|
0.285
|
-0.537
|
0.501
|
4
|
-0.152
|
0.093
|
-0.498
|
0.141
|
5
|
-0.178
|
0.079
|
-0.440
|
-0.111
|
6
|
-0.131
|
0.088
|
-0.431
|
-0.293
|
SD error
|
0.105
|
0.058
|
0.019
|
0.147
|