Improving the Performance of Vegetable Leaf Wetness Duration Models in Greenhouses Using Decision Tree Learning

Abstract

Leaf wetness duration (LWD) is a key driving variable for peat and disease control in greenhouse management, and depends upon irrigation, rainfall, and dewfall. However, LWD measurement is often replaced by its estimation from other meteorological variables, with associated uncertainty due to the modelling approach used and its calibration. This study uses the decision learning tree method (DLT) for calibrating four LWD models-RH threshold model (RHM), the dew parameterization model (DPM), the classification and regression tree model (CART) and the neural network model (NNM)-whose performances in reproducing measured data are assessed using a large dataset. The relative importance of input variables in contributing to LWD estimation is also computed for the models tested. The LWD models were evaluated at two different greenhouse locations: in a Chinese (CN) greenhouse over three planting seasons (April 2014-October 2015 and in a Spanish (ES) greenhouse over four planting seasons (April 2016-February 2018). Based on multi-evaluation indicators, the models were given a ranking for their assessment capabilities during calibration (in the Spanish greenhouse from . The models were then evaluated on an independent set of data, and the obtained areas under the receiver operating characteristic curve (AUC) of the LWD models were over 0.73. Therein, the best LWD model in this case was the NNM, with positive predict values (PPVs) of 0.82 (SP) and 0.90 (CN), and mean absolute errors (MAEs) of 1.85 h (SP) and 1.30 h (CN). Consequently, the DLT can decrease LWD estimation error by calibrating the model threshold and choosing black box model input variables. simulation were developed, and are routinely used to provide inputs to early disease warning systems of disease outbreaks [6,7]. The LWD models can be divided into two broad categories: empirical and physical models. The latter are data demanding because they require inputs that are not always available, such as cloud cover and net radiation [8,9]. Conversely, estimation of LWD is based on its relationships with meteorological variables available in standard agro-meteorological stations. Some examples of empirical LWD models are the simple relative humidity threshold model (RHM), which simulates the leaf wetness occurrence when humidity is above a threshold [10,11]; the dew point depression method (DPM), based on the principle of dew formation [12,13]; and the classification and regression tree (CART) model, which considers the non-linear relationship between leaf wetness, wind speed, rainfall, dew temperature and relative humidity in decision nodes to determine leaf wetness [14]. Kim et al. demonstrated the spatial portability of the CART model through its application in different environments [15]. Gillespie et al. achieved good results when estimating LWD by a dew point depression model [16]. Further attempts have been made for estimation of LWD using neural network models (NNMs), which can be considered black box models. These are self-adaption models trained with refe...

Dew Yield and Its Influencing Factors at the Western Edge of Gurbantunggut Desert, China

“…Although many models based on energy balance can estimate dew production [63][64][65][66][67], some parameters of the models, such as soil heat flux, heat exchange coefficient and the properties of the condensing surface, are complex and difficult to obtain. Thus, a simplified model that would demand only a few regular, simple data, available everywhere in the world, would be much more helpful.…”

The Use of Artificial Neural Networks for Forecasting of Air Temperature inside a Heated Foil Tunnel

“…As the greenhouse system is nonlinear and variable in time, ANNs have been applied to greenhouse environment modeling, including internal temperature. For the first time, ANNs were used to predict greenhouse temperatures in the early 1990s [24,57,58]. Ferriera at al.…”

“…Modelling the microclimatic parameters of a greenhouse is critical to optimize internal climatic conditions of greenhouses during different stages of plant growth. The basic problem is to develop an accurate model [7,19,23,24].The greenhouse environment is a very complex, nonlinear, multi-input and multi-output dynamic system. It is a complex thermodynamic system-the internal greenhouse climate is a function of heat and mass transfer.…”

“…There are also time delays and intensive disturbances from the surroundings. This makes establishing an accurate mathematical model of a greenhouse's microclimate intricate and difficult [2,4,6,18,20,24].A great deal of research has been done in greenhouse modeling in the recent years. Many models have been developed to predict energy consumption and greenhouse microclimatic parameters.…”

“…Modelling the microclimatic parameters of a greenhouse is critical to optimize internal climatic conditions of greenhouses during different stages of plant growth. The basic problem is to develop an accurate model [7,19,23,24].…”