Predicting the date of harvest of vining peas by means of air and soil temperature sums and node countings

Abstract

Predictions of the date of harvest of vining peas ( Pisum sativum L.) were performed by means of statistical models. Simple models based on sowing-to-harvest prediction criteria and two-component models based on a combination of sowing-to-emergence and emergence-to-harvest prediction criteria were applied. The criteria used were number of days, and soil and air temperature sums, with base temperature in the range 2.6–3.8°C and 3.0–6.0°C, respectively. The best predictions of the harvest date were obtained with two-component models, and sowing-to-emergence predictions criteria based on soil temperature were better than others examined. Among two-component models, the smallest error of prediction-2.4 days-was obtained with the model using sowing-to-emergence soil temperature sums with base temperature 3.0°C, and emergence-to-harvest number of days as prediction criteria. The prediction model traditionally used in practice applies sowing-to-harvest air temperature sums with base temperature 4.5°C as prediction criterion, and the error of prediction of this model was 4.3 days. Close linear relationships were found between number of nodes and air temperature sums accumulated from sowing or emergence. In consequence, the number of nodes is a reliable indicator of the actual air temperature sum status of the crop. Some practical aspects of applying two-component models and node counting for predicting the date of harvest of vining peas are discussed.