Evaluation of the ‘Fertigation Model’, a Decision Support System for Water and Nutrient Supply for Soil Grown Greenhouse Crops

Soil grown greenhouse crops require high fertilisation rates. Combined with the common practice of over-irrigation, leaching of nutrients is a serious problem. In order to reduce the environmental impact, a ‘fertigation’ model was developed as a decision support system for irrigation and fertiliser supply. The applicability in growers practice was evaluated during two years on commercial nurseries, growing chrysanthemum (Dendranthema grandiflorum). The evaluation was performed by comparison of the actual water and irrigation strategy of the growers with the strategy recommended by the model in a specific section within the same greenhouse. At one chrysanthemum grower a lysimeter was installed to measure water and nutrient leaching. The model performed well in general, without any yield or quality decline by using the model. The irrigation and the nitrogen surplus were decreased significantly compared to the growers standard and consequently reduced the environmental impact. The results indicate also that application of this model depends highly on the growers’ attitude towards the environmental impact of irrigation and fertilisation at one hand and the avoidance of risks at the other. INTRODUCTION In contradiction with substrate grown greenhouse crops, soil-grown greenhouse crops in the Netherlands contribute to the pollution by nutrients (N and P) of surface and groundwater substantially (Wunderink, 1996; Boers, 1986). The main reason is the open character of the root environment, with either free leaching of the irrigation surplus to ground water or to drainage systems with discharge directly to surface waters. Drainage water is (partly) re-used at some nurseries but is not a solution in many cases (Voogt and Korsten, 1995). The main cause of the important nutrient emission is the high fertiliser supply, not only for the crop demand but also to achieve high nutrient levels in the root environment. Combined with the common practice of over-irrigation and the fact that in the Dutch soil-grown greenhouse crops fertigation is the standard for fertilisation, leaching of nutrients is a serious problem (Sonneveld, 1995). An apparent solution is to bring down the over-irrigation, to avoid or at least reduce leaching. For this purpose a socalled ‘fertigation’ model was developed. This is a decision support system for irrigation and fertiliser supply, aiming at supply in match with the crop demand (Voogt et al., 2000). The objective of this work is to investigate the applicability in growers practice. Chrysanthemum was chosen because this crop is the major soil grown greenhouse crop in the Netherlands, moreover the crop at which the nutrient emission appeared to be severe (Baltus and Volker Verboom, 2005). The evaluation was performed by assessing the water and irrigation strategy of three chrysanthemum growers during two years, comparing the growers’ own strategies with those recommended by the model in a specific section within the same greenhouse. At one nursery a lysimeter was installed to measure water and nutrient leaching. Specific attention was paid to the applicability of the model in the grower’s perception of irrigation and fertilization management for their specific growth concept. Proc. III IS on HORTIMODEL2006 Eds. L.F.M. Marcelis et al. Acta Hort. 718, ISHS 2006 532 MATERIALS AND METHODS Three year-round chrysanthemum nurseries, situated in a river clay polder in the centre of the Netherlands, each in the vicinity of 10 km from each other were under investigation. The greenhouses were modern and 7, 2 and 1 year old, with a light transmission of 72, 79 and 80% for grower A, B and C respectively. All greenhouses were equipped with supplementary lighting. All three were more or less heavy clay soils, with 21 till 28% clay (< μm) respectively and a ground water level between 80 and 90 cm below field level. Treatments Two treatments: Standard and Model were compared at all growers, in adjoining irrigation sections of the greenhouses, both sections differed at maximum only 3 days in planting and the average harvesting date. Standard, the growers’ own strategy on irrigation and fertilization was used; Model, irrigation and fertilization was applied according to the model calculations, with the following restrictions: irrigation frequency was performed according to the growers’ judgement, for decisions on irrigation quantity and nutrient concentrations the output of the model was leading, however the grower was free to adjust the recommended supply to his own judgment. The fertigation model used is an empirical model and consists of separate algorithms for the estimation of evaporation and nutrient uptake (Voogt et al., 2000). As a consequence of the incomparability of the three nurseries (time, place, growing conditions), no statistical analysis were performed on the data. The yield results of each individual crop were analysed by ANOVA. Evapotranspiration Model A linear regression model is used based on variables routinely recorded by climate computers (de Graaf and van den Ende, 1981; de Graaf, 1988). The total evapotranspiration integrated over time in mm is defined as Tc = (at Ro + bt DM) s (1) where Ro = global radiation measured outside the greenhouse integrated over time (J cm -2 ), DM = temperature difference between the heating pipes and the greenhouse air temperature in °C, integrated over time in ‘degree minutes’, at, bt = empirical crop factor: 2.0*10 -3 (mm (J cm -2 ) -1 ) and 0.18*10 -4 (mm (°C min) -1 ) for factor at and bt respectively for chrysanthemum, s = plant size factor, defined as the ratio of the actual plant length (cm) to mature plants (LAI > 3). Equation (1) was modified to the situation in modern greenhouses, with artificial lighting, to introduce the effect of light transmission in and the effect of screening: Tc = (at (c st Ro + Ra) + bt DM) s (2) where Ra = effective radiation from supplementary lighting during operating hours, c = factor for light transmission of the greenhouse, expressed as the ratio of the actual transmission to 0.68 (being the factor of the greenhouses the empirical crop factors a and b were derived from) and st = factor for opening (1) or closure (0) of the screen. During the cropping period, Tc was computed continuously by the climate computers. The necessary driving variables were measured every minute; global radiation was measured with solarimeters; Platinum resistance Thermometers and NTC sensors were used to measure temperature from the greenhouse air and the heating pipes; plant length was determined by an empirical growth curve for chrysanthemum and was weekly checked by observations. Nutrient Uptake A simple approach was used, given the fact that nutrient uptake is closely related to the water uptake and that the soil buffer is very large in relation to the daily uptake