Temperature monitoring strategy for microclimate prediction in low-automation greenhouses: a preliminary analysis

Climate control strategies in low-automation greenhouses typically rely on the measurements from a single sensor. Indeed, implementing more complex monitoring devices and sensor networks may increase investment costs without necessarily improving profitability. This paper presents a low cost and low complexity temperature monitoring strategy, which is illustrated in a case study of tomato cultivation during the mid-season in a traditional Mediterranean greenhouse located in Pisa, Italy. The objective is to evaluate the temperature distribution within a portion of the greenhouse and to investigate the representativeness of single-point measurements for assessing the local microclimate. During the experimental campaign, the vertical and horizontal temperature differences reached maximum values of 9.9 °C and 7.3 °C, respectively. The performed temperature measurements appear to be correlated (R2 > 0.95), and this information was exploited for the prediction of the greenhouse temperature in multiple points, resulting in an average RMSE of 1.3 °C, with differences depending on the specific position taken as reference. These findings offer insights into the representativeness of single-point measurements and the optimal positioning of the sensor station within the greenhouse. The inclusion of microclimate heterogeneity in climate control strategies can help minimise the local presence of unfavourable growth conditions, the excess of energy use, and installation and maintenance costs.