Delayed harvesting of bananas with 'sealed' covers on bunches. 1. Modified atmosphere and microclimate inside sealed covers

A field study was conducted to examine the effects of various bunch covering and harvest delay combinations on the environment inside bunch covers, and on fruit yield and quality parameters. Treatments included unsealed and initially sealed (but often ultimately leaky) clear covers under the industry standard blue/silver cover, and the inclusion of an ethylene absorbent inside the 'sealed' covers. This paper reports on the microclimate inside the bunch covers, and the modified atmosphere that developed inside the 'sealed' covers. During clear weather in January, average air temperatures inside bunch covers were 3, 5, and 6C above ambient for unsealed single covers, for unsealed double covers, and for 'sealed' double covers respectively. Differences were smallest at dawn, and greatest in the late afternoon. During clear weather in July bunch cover temperatures were 1, 1.5, and 2C above ambient respectively. Relative humidities were also influenced by covering, with 'sealed' cover humidity remaining above 95%. Carbon dioxide concentrations inside 'sealed' covers when bunches were harvested averaged 8%, but varied widely from <1 to 42%. Highest CO2 concentrations were obtained when weekly mean air temperatures were greater than 22C. Oxygen concentrations averaged 10% (range 2-19%), were lowest during the warmer part of the year, and were inversely related to CO2 levels. Ethylene concentrations varied between 0.03 and 9 L/L, were highest during the hotter part of the year, and were reduced by about 60% with the inclusion of KMnO4 inside 'sealed' covers. An iterative model with a .l h time step was used to simulate the expected O2 and CO2 levels inside the 'sealed' covers. Simulation results indicated that polyethylene permeability alone could not balance the computed respiratory gas exchange, and account for the changes in O2 and CO2 concentrations. It was necessary to incorporate a substantial leakage component into the model in order to simulate successfully the field observations.