A relook at the epidemiology of Cercospora spot on avocado in South Africa.

Abstract

Avocado (Persea americana Mill.) belongs to the family Lauraceae and is one of the most economically important subtropical fruit crops in the world. The South African avocado industry contributed approximately “R1.2 billion to the total gross value of subtropical fruits (R3.4 billion) during the 2017/18 season”, according to the latest available records. One of the most serious pre-harvest diseases affecting avocado in South Africa is Cercospora spot. Losses of up to 70% have been reported on unsprayed trees. This disease is commonly found in avocado producing regions where warm, humid and rainy conditions persist. It affects all commercial cultivars, with ‘Fuerte’ being recognized as the most susceptible cultivar. The plant pathogen responsible for this disease is Pseudocercospora purpurea (Cooke) Deighton. As with other Cercospora species, this fungus grows slowly and sporulate sparsely on artificial media. Typical disease symptoms are found on the leaves, stems and fruit. Lesions first appear on the underside surface of leaves. These lesions are minute and are brown in colour. As the disease progresses, lesions are observed on both sides of leaves and have distinctive chlorotic halos. On the fruit, small lesions form, later becoming sunken, irregular and brown to black in colour. The most commonly chemical control is copper oxychloride although there are other registered fungicides for use against Cercospora spot. The South African Avocado industry currently uses a predictive model, developed by Dr J.M. Darvas, in the early 1980s, to predict the number of conidia and the timing for the first spray. The model is based on the temperature and rainfall that occurred in the week preceding the calculation of the prediction. adaptions in the fungal populations over the years, it was vital to re-evaluate this model. The primary aims of this study were to determine whether the current Darvas 2 model is still valid for forecasting the first spray for effective control; secondly, to evaluate whether the inclusion of humidity and/or leaf wetness values into a model would enhance its predictive accuracy; and thirdly, to evaluate the size of fruit that was susceptible to infection by P. purpurea. In this study, spore trapping and critical infection trials (bagging trial) were conducted for two seasons. Spore traps were placed in two unsprayed ‘Fuerte’ orchards (HL Hall and Sons and the ARC-TSC) in the first season (2017/18). However, in the second season (2018/19), only one orchard (ARC-TSC) was used for both trials because no conidia were trapped as a result of a low disease incidence at the Halls orchard that was used in Season One. At harvest, fruit were assessed for Cercospora spot using a disease rating scale. The disease index data for both seasons (2017/18 and 2018/19) were correlated with weather data using multiple stepwise linear regression analysis. In both seasons, the critical infection period was in the beginning of the season. It was also established that fruit exposed to natural infection early in the season from October to November developed significantly more Cercospora disease symptoms than fruit exposed later in the season. The daily spore trapping results (2017/18 season) indicated that conidia were mostly trapped on days when rainfall occurred. The most significant correlation (r=0.893) was found between the weekly number of trapped conidia and weekly rainfall (September to December 2017). Based on the weekly spore trapping results of the 2018/19 season, for the period October to December, there was a strong correlation (r=-0.696) between conidia and mean maximum temperature. For the entire season (October to April 2018) the correlation between conidia and mean maximum temperature was slightly lower (r=-0.520). In the 2017/18 season, more rainfall fell and more conidia were trapped than in the 2018/19 season. Due to low rainfall during the 2018/19 season, a stronger correlation was found between conidia and temperature than conidia and rainfall. This negative correlation can be explained by the cooling effect of rain, as mentioned by Darvas (1982). For both seasons, the weekly weather parameters and the weekly spore trapping data were correlated with one another. Using multiple stepwise linear regression analysis of the weekly conidia trapped and weekly weather data, three models were developed for each season. It was found that all new models (for each season) followed a similar pattern to the Darvas 2 model, with some minor differences. The spore trapping results confirmed that rainfall and temperature were the dominant environmental parameters. However, leaf wetness and relative humidity were not factors in the release of conidia but played a role in disease development, probably in the step of host infection. The study found that the Darvas 2 model was still an effective forecasting tool. However, the selected model/s (current Darvas 2 model or the new models) must be used in combination with fruit size monitoring to determine accurate and cost-effective timing of the first spray. This study determined that the first spray should be applied around mid-October (depending on the geographic region, rainfall and Z values). In addition, it was also concluded that spraying should begin when the Z-value is 15, and fruit size is approximately 25mm in diameter, and not 40mm as previously recommended by Darvas (1982). This study showed that spraying when the fruit is 40mm in diameter would be too late to slow down disease development. In support of the primary aims, experiments were conducted to determine the growth requirement/s of P. purpurea. The growth of P. purpurea was evaluated on several artificial media (potato dextrose agar (self-made), potato dextrose agar (commercial), malt extract agar, potato sucrose agar, oatmeal agar (self-made), oatmeal agar (commercial), and V8 juice agar. The fungus was grown on these media at temperatures ranging from 5oC to 35oC. The radial growth was recorded by measuring the colony diameter for a period of 28 days at seven-day intervals. The results of the growth study indicated that oatmeal agar was the best agar medium, and that 25oC was the optimal temperature for the growth of P. purpurea on artificial media.

In conclusion, this study showed that the Darvas 2 model is still an effective forecasting tool, irrespective of climate change and that when the model i.e., either the current Darvas 2 model or one of the newer models is used in combination with fruit size monitoring, we can achieve a more accurate and cost-effective time to apply the first fungicide spray.