doi:10.5455/vetworld.2013.321-324 Multiple linear regression model for forecasting Bluetongue disease outbreak in sheep of North-west agroclimatic zone of Tamil Nadu, India

Aim: A study was undertaken to develop a forecasting model for predicting bluetongue outbreaks in North-west agroclimatic zone of Tamil Nadu, India. Materials and Methods: Eleven bluetongue outbreaks were characterised by active and passive surveillances for a period of twelve years and used in this study. Meteorological data comprising of maximum and minimum temperatures, relative humidity, rainfall and wind speed were collected and used as the multiple predictor variables in the multiple liner regression model. Results: A multiple liner regression model was developed for the North-west zone of Tamil Nadu. Values of the dependant variables were less than or greater than one, and indicated remote or greater chances of bluetongue outbreaks respectively. The monthly mean maximum and minimum temperatures, relative humidity at 8.30 h and at 17.00 h IST, wind speed, and monthly total rainfall of 29.1 - 31.0°C, 20.1 - 22.0°C, 80.1 - 85.0%, 65.1 - 70.0%, 3.1 - 5.0 km/h and < 200 mm respectively, were identified as the ideal climatic conditions for increased numbers of bluetongue outbreaks in this zone. Conclusion: Based on the values obtained from the prediction model, stake holders can be warned timely through the media to institute suitable prophylactic measures against bluetongue, to avoid economic losses due to disease.


Introduction
vectors.Culicoides vectors are significantly affected by the climate and weather and frequently influence the Bluetongue (BT), an infectious and nonincidence and overall severity of BT [6].BTV may be contagious disease of ruminants, is caused by BT virus introduced to new regions by the movement of infected (BTV), an RNA virus which belongs to the Orbivirus animals, but will survive in a new region only if genus of the family Reoviridae [1].The disease is competent vectors and sufficient susceptible hosts are characterized by mortality rates as high as 70% in present.Environmental temperature [8], relative highly susceptible sheep populations [2].The humidity [9], effective rainfall [10,11], wind direction estimated annual economic loss due to BT outbreak and wind speed [12] have been implicated in the natural was Rs. 52 lakhs in TamilNadu and about $ 3 billion spread of infected culicoides vectors.worldwide [3,4].Bluetongue is endemic in Indian Timely announcement of weather forecasting is a states of Tamil Nadu, Andhra Pradesh, Karnataka, useful tool for sheep owners [5] to carry out disease Maharashtra, Gujarat, Rajasthan, Haryana, Himachal control activities to curtail the outbreaks and eliminate Pradesh and Jammu and Kashmir.In Tamil Nadu, 22 the disease and infection in the shortest possible time out of 23 districts were reported to be affected by the frame, using the most cost-effective ways.Hence, the BTV [5].More than 27 Culicoides species have been research work was carried out to develop an effective identified in India, and Culicoides imicola, C.
BT forecasting system using maximum and minimum peregrinus, C. oxystoma and C. brevitasis are the temperature, relative humidity, rainfall and wind speed predominant species involved in the transmission of in North-west agroclimatic zone of Tamil Nadu, India.BT in Tamil Nadu [5,6,7].
The occurrence of BT depends on timing and   were more outbreaks with monthly total rainfall ranges Relative humidity at 17.00 h IST(per cent): 42.33 -67.83 of 0 -100 mm, followed by 100 -200 mm (Table 2).Monthly total rainfall (mm): 1.31 -409.20This indicated that low to moderate levels of rainfall Monthly mean wind speed (km/h): 1.33 -8.20 are sufficient for the multiplication of vectors, and that heavy rains inundate and erode the breeding sites as Discussion earlier observed by Braverman [12].Ward [20] also Forecasting model: In this model, preceding 30 day reported that rainfall of two to three inches (50 to 75 averages of the predictor variables are used to give mm) was sufficient to allow the vector species and early warning on a daily basis.The model is Subramanian and Piramanayagam [10] opined that qualitatively valid only for the study zone within the rainfall prevailing during the months of November working range.If the value of a dependant variable is (148 mm) and December (176 mm) predisposes high less than one (y<1), there is a remote chance for a morbidity and mortality rate in Tirunelveli region of disease outbreak; and if greater than one (y>1), there is Tamil Nadu, India.an increased chance of disease outbreaks as previously Monthly mean wind speed vs. bluetongue outbreaks: reported by Jong [15].It is concluded that based on Increased numbers of outbreaks were observed with results of the prediction model, stake holders may be monthly mean wind speed ranges of 3.1 to 5.0 km/h continuously advised about possible future bluetongue when compared to lower or higher wind speed ranges outbreaks through the media, for instituting suitable (Table -2).This is in accordance with the findings of prophylactic measures to avoid economic losses due to Ausvetplan [21] which reported that the lower wind the disease.
speed encouraged local spread, as the insects would not Correlation of meteorological data with outbreaks fly in higher wind speeds (>8 km/h).Moreover, low Monthly mean temperature vs. bluetongue outbreaks: wind speeds favour vector seeking of hosts for blood In this zone, BT outbreaks were increased when the meal feeding, mating, egg laying and shelter as stated monthly mean maximum and minimum temperatures by Sellers [22].Hence, wind can aid the passive ranged from 29.1 to 31.0°C and 20.1 to 22.0°C dispersal of infected vector as reported by Saegerman respectively (Table 1  read and approved the final manuscript. 12. Braverman, Y., Chechik, F. and Mullens, B. (2001).The interaction between climatic factors and bluetongue GS and AB designed the study.GS conducted the S., Vanzetti, T., Groit, C. and Stark, K.D.C. (2007).Use of study and analyzed the data.DR, DK and MG mapping and statistical modeling for the prediction of bluetongue occurrence in Switzerland based on vector drafted and revised the manuscript.All authors biology.Vet.Ital., 43: 513-518.

Table - 1
:10.5455/vetworld.2013.321-324 .Correlation of the monthly mean maximum and minimum temperatures and relative humidity at 8.30 h IST, with the number of bluetongue outbreaks recorded nine suspected flocks through active surveillance for a ture, minimum temperature, relative humidity (at 8.30 period of two years (June 2007 to May 2009).Based on h IST), relative humidity (at 17.00 h IST), wind speed the history and clinical symptoms in affected sheep, and total rainfall correspondingly.Multiple linear whole unclotted (EDTA) blood was collected as per regression model used as follows; standard protocol prescribed by CFSPH [2] and y=β +β w +β w +β w +β w +β w + β w + β w 0 Virology, Indian Veterinary Research Institute, β -Unstandardised coefficient for each predictor (1-6) Mukteswar, Uttranchal for confirmation of bluetongue.variables, A total of six outbreaks were identified through w -Monthly mean maximum temperature (°C), 1 passive surveillance for a period of twelve years (June w -Monthly mean minimum temperature (°C), 2 1997 to May 2009) from data available at Animal w -Relative humidity at 8.30 h IST (percent), 3 Disease Investigation Units, Salem and Dharmapuri, w -Relative humidity at 17.00 h IST(percent), 4 Tamil Nadu.w -Monthly total rainfall (mm), 5 Meteorological data: Meteorological data comprising w -Monthly mean wind speed (km/h), 6 monthly mean i).maximum temperature; ii).minimum w tow -Predictor variables y = (220.453)+ (-8.922 w ) + ( 5.358 w ) + (0.183 w ) + logical Department, Chennai, Tamil Nadu, and used in this study.Tables-1 and 2 indicate the number of Model summary and regression co-efficient are outbreaks experienced and values of the predictor shown in Table-3.The model was developed following variables.predictor variables working ranges and fixed based on the climatic conditions that prevailed over the past 12 Statistical analysis: Statistical Package for Social doi

Table - 2
. Comparison of monthly mean relative humidy at 17.00 h IST, total rainfall (mm) and wind speed (km/h), with the number of bluetongue outbreaks experienced

Table - 3
High relative humidity was found to be a favourable factor for the development of Culicoides outbreaks and a multiple linear regression model was species [19] and it can positively alter the level of developed and validated with available data.This is a activity of Culicoides midges [9].Similarly, more pioneer model for the North-west agroclimatic zone of outbreaks were recorded during high monthly mean Tamil Nadu, India.Based on the value obtained from .Model summary and regression coefficient for the developed model Indian J. Vet.Pathol., 32: 111-124.the prediction model, stake holders can be continuously 9. Wittmann, E.J. and Baylis, M. (2000).Climate changes: advised through the media to institute suitable Effects on Culicoides transmitted viruses and implications prophylactic measures to avoid economic losses due to for UK.Vet.J., 160: 107-117.bluetongue.10.Subramanian, K.S. and Piramanayagam, S. (2001).Epidemiological observation on bluetongue in Tirunelveli Authors' contribution region.Indian Vet.J., 78: 945-946.11.Racloz, V., Presi, P., Vounatsou, P., Schwermer, H., Casati, Monthly mean relative humidity vs. bluetongue meteorological parameters were correlated with BT outbreaks: