Weather based forewarning models for coffee borer incidence

Coffee berry borer(CBB) and shot hole borer are the two pests that damage coffee berries to a considerable extent in a favourable weather environment. The coffee berry borer can cause bean yield losses of 30–35% with 100% of perforated berries at harvest time; nevertheless, damage can be greater if harvest is delayed. All the commercial coffee varieties and species are attacked by this insect. The coffee growing areas of Niligiri district of Tamil Nadu, Wayanad district of Kerala and Kodagu district of Karnataka together form a contiguous coffee growing tract. The CBB infestation in this region had spread from the Niligiris towards the north covering Wayanad and south Kodagu over a period of four to five years(Anonymous, 2002). The shot-hole borer, which attack green succulent branches of robusta and green scales, which is a sucking pest of both arabica and robusta are major pests of coffee. Although they are major pets, they are not in serious nature except in some endemic areas. Therefore, an attempt has been made in this paper to investigate the statistical relationship exists between coffee berry borer and shot hole borer infestation and meteorological factors and also to develop forewarning models for predicting the population of these pests.

A study was conducted in Chundale, Wayanad district in Kerala, to assess the seasonal incidence and influence of weather on coffee berry borer, Hypothenemus hampei(Ferrari) and shot hole borer(Scolytus rugulosus) from January 2008 to December 2011 in the 30 years old Robusta coffee plantations (variety S.274 and CXR) at Regional Coffee Research Station, Chundale. The experimental plots consist of one acre plantation and 20 plants were selected for recording pest data at fortnightly interval. Observations were made from 20 randomly selected coffee plants and in each plant, three branches at top, middle and bottom were chosen to record the pest data. The damaged berries by berry borer were collected and adult beetle populations were counted. The number of adult beetle populations in shoots were counted to record population of shot hole borer. The data on weather parameters, maximum temperature (°C), minimum temperature (°C), relative humidity (%) and rainfall (mm) were recorded from Automatic Weather Station installed in the station during the study period. The fortnightly mean of pest populations of 20 plants were calculated from the data on the incidence of coffee berry borer and shot hole borer and these were correlated with weather parameters.
Correlation coefficient (CC) between pest population and weather parameters has been worked out for fortnight when maximum infestation was observed. Student's 't' test (Fisher and Yates, 1938) was applied to test the significance. The fortnight's individual weather parameters having highest correlation coefficients which are statistically significant were selected. All these statistically significant parameters were subjected to multiple correlation and stepwise regression combining all the selected parameters for these selected fortnights and a 'F' test was performed for testing its significance. Regression equations were developed for the fortnight with maximum infestation using the significant weather parameters having highest correlation coefficients. Graphical analysis between significant weather parameters and pest population were made.
The coffee berry borer attack was mainly confined between first fortnight of January and second fortnight of April in both the varieties. The peak population was observed during first and second fortnight of April which correspond to fruit set of coffee in Wayanad (Table 1). From first fortnight of June with the onset of rainfall, the population of CBB decreased in the present study. Rainfall was possibly the factor that attributed to the low berry borer population (Ferreira et al., 2000).  A critical examination of various weather parameters showed that CBB population increases substantially when the maximum temperature was above 31.5 0 C and minimum temperature was above 19.5 0 C and rainfall below 21.1mm. Jaramillo et al. (2009) reported lower threshold values of temperature, lower optimal Populations of SHB were observed throughout the year and peak population of the pest was observed between first fortnight of November and 2 nd fortnight of January. The maximum attack was observed at 2 nd fortnight of January followed by first fortnight of January which corresponds to budding stage of coffee in Wayanad. It is also a serious pest of Robusta attack on tertiary twigs, young primaries and suckers. Incidence start by mid-June temperature, upper optimal temperature and upper threshold temperature for population growth of CBB are 14.9, 23, 30 and 32 0 C respectively. These critical limits of weather variables were mostly prevailed in maximum population year 2009. The maximum temperature and minimum temperature during previous four fortnights showed increasing trend towards the maximum infestation week at 2 nd fortnight of April in 2009. Rainfall showed decreasing trend towards the maximum infestation at 2 nd fortnight of April during 2009. Shot hole borer population increases substantially when minimum temperature was below 17.5 0 C, mean RH was below 78% and rainfall above 11.4mm. These critical limits of weather variables were mostly prevailed in maximum population year 2011. Minimum temperature and mean RH showed decreasing trend towards the maximum infestation at 2 nd fortnight of April during 2011. Rainfall showed decreasing trend towards the maximum infestation at 2 nd fortnight of January during 2011.
Correlation coefficients worked out between fortnightly adult beetle counts of CBB and different weather parameters (Table 1). Correlation studies indicated that different weather parameters influenced the pest population differently. In the present study population of CBB at 2 nd fortnight of April was found to have highest positive correlation with maximum temperature and minimum temperature of 2 nd fortnight of April. Rainfall of 2 nd fortnight of April have highest negative correlation with pest population of 2 nd fortnight of April. Relative humidity was not significantly correlated with CBB incidence. Minimum temperature during 2 nd fortnight of October have highest negative correlation with SHB incidence of 2 nd fortnight of January. Relative humidity during 2 nd fortnight of November have highest negative correlation with SHB incidence of 2 nd fortnight of January. Rainfall during first fortnight of January have highest positive correlation with SHB incidence of 2 nd fortnight of January.
The multiple regression equation which describes the average relationship between the CBB population and Y= CBB Population at 2 nd fortnight of April. The multiple correlation coefficient was 0.98 which was significant. The results of the study indicates the population of CBB may be predicted with the current fortnight's weather data.
The multiple regression equation which describes the average relationship between the SHB population and significant weather parameters is derived and expressed as below: Y= -622.960 + 7.620 RH_Nov.II + 0.656 RF_Jan.I (R 2 = 0.90) Y= SHB Population at 2 nd fortnight of January. The multiple correlation coefficient was 0.95 which was significant. The model indicates the population of SHB may be predicted with the with mean RH data of four fortnights and rainfall data of previous fortnight of the incidence of pest.