Document donwnloaded from:
[ http://redivia.gva.es/handle/20.500.11939/4555 ]
This paper must be cited as:
[Suay, R., Martínez, P.F., Roca, D., Martinez, M., Herrero, J.M., Ramos, C. (2003). Measurement
and estimation of transpiration of a soilless rose crop and application to irrigation management.
Proceedings of the Sixth International Symposium on Protected Cultivation in Mild Winter Climate:
Product and Process Innovation, Vols 1 and 2, (614), 625-630.]
The final publication is available at
[https://doi.org/10.17660/ActaHortic.2003.614.93]
Copyright [ISHS]
MEASUREMENT AND ESTIMATION OF TRANSPIRATION OF A
SOILLESS ROSE CROP AND APPLICATION TO IRRIGATION
MANAGEMENT
R. Suay
1
, P.F. Martínez
1
, D. Roca
1
, J.M. Herrero
2
, C. Ramos
2
1
Instituto Valenciano de Investigaciones Agrarias (IVIA). Apdo. Oficial. 46113 Moncada,
Valencia, Spain. (rsuay, pfmarti, droca)@ivia.es
2
Dept. Ingeniería de Sistemas y Automática. Universidad Politécnica de Valencia. Apdo.
de Correos 22012 Valencia, Spain. (juaherdu,cramos@isa.upv.es)
Additional index words: hydroponic culture, modelling, leaf area index, closed system.
Abstract
Drained nutrient solution of soilless culture crops in many countries, is
rejected to waste, producing undesirable polluting effects and reducing the water
and fertilizer use effciency of the cropping system.
Current irrigation practices in protected soilless culture, are based either on
fixed interval scheduling, corrected by means of EC monitoring in drainage or
substrate extract, or on crop transpiration estimations by means of solar radiation
integration.
Closed soilless culture systems help to improve water and fertilizer use
efficiency, but irrigation management needs to be optimized, and properly operated
in relation to crop requirements, and needed proportions of air and water in the
substrate.
A study has been carried out, where the transpiration of a heated rose crop,
conducted by the shoot bending system, is continuously measured by means of an
electronic balance, and estimations are calculated with a simplified Penman-
Monteith equation, adjusted by weekly LAI estimations, as well as outside solar
radiation and inside water vapour saturation deficit recordings.
Based on these transpiration estimations, an irrigation strategy has been
tested and compared to current irrigation practices used in commercial holdings.
1. Introduction
During recent years, a number of authors are paying attention to the improvement
of irrigation of protected crops. The control equipment for irrigation has improved
considerably, but growers of the mediterranean regions are mostly managing irrigation on
the basis of empirical criteria and rules. Fixed volumes of water supply at regular time
intervals, or a partial use of solar radiation integration, is the most common way of
automatic irrigation. Microclimate in mild winter greenhouses can be a strong source of
water stress for crops and the risk is increased by the fact that soilless culture has very
much increased, and substrate physical properties of low water retention force the grower
to look for a reliable automatic system based on optimum decision criteria. Otherwise
either the crop can be under water stress danger or the system can be working on the basis
of extra expenses of water and fertilisers.
More or less complex models of transpiration based on the Penman-Monteith
equation (Monteith, 1973) have been proposed and validated under different conditions
(Bailey et al., 1993; Gonzalez, 1995; Lorenzo et al., 1998; Medrano et al., 2001). A
simplified version (Stanghellini, 1987; Boulard and Jemaa, 1993; Baille et al., 1994a;
Medrano, 1999) that takes into account both, solar radiation and air saturation deficit as
variables could be very useful for the implementation of irrigation algorithms in a soilless
culture system.
The present paper aims at adjusting and validating a simple model of crop
transpiration for a rose soilless crop. A model-based algorithm has been integrated in a
control system, and has been tested for the complete management of crop irrigation needs,
in order to check the real possibilities and degree of reliability for practical applications
2. Materials and methods
A soilless rose crop (cv. Dallas) is grown in an acrylic covered greenhouse of 250
m
2
. Heating is supplied for a minimum of 16 ºC, by means of air forced aerotherms, and
humidity is kept at 50 % minimum with a high pressure fog system. Two 30 plants units,
located at the central position of the greenhouse, are grown in a NFT type system, where
transpiration is measured at short time intervals (15 seconds), by means of an electronic
balance (resolution ± 0.1 g). Nine hundred plants grown in a perlite hydroponic system
complete the rest of the greenhouse. All plants are grown following the local standard
commercial technics, using the stem bending technique and all-year-round production.
Short time soilless rose crop transpiration has been recorded and compared with a
simplified transpiration model during a winter cropping period from November to
February. A simplified transpiration model based on the simplified Penman-Monteith
equation has been derived:
Where:
E = crop transpiration rate (g m
-2
h
-1
)
G = outside solar radiation, w m
-2
VPD = inside air vapour pressure deficit, kPa
k = radiation extinction coefficient, 0.64
A,B = equation parameters (A dimensionless, B g m
-2
h
-1
kPa
-1
)
This simplified model considers that the transpiration is mainly explained by two
components or terms: a radiative part (directly related to the radiation G absorbed by the
crop) and an advective part (directly related to the inside greenhouse vapour pressure
deficit). The advective term has shown to be really important, especially in winter nights
when heating is working on with no radiation or in cloudy days.
In order to implement a real-time irrigation management, short time (15 seconds)
estimations of crop transpiration have been used. A and B coefficients have been
calculated by statistical identification (multi-parameter lineal regression or multiple
regression model) for transpiration versus radiation and vapour pressure deficit real-time
records. Taking into account that commercial greenhouses may have an outside weather
station and an inside psychrometer, the models obtained by both, outside and inside solar
radiation have been compared, finding that there is no difference in the fitting of their
prediction (R
2
= 0.92 for both). Thus, the model with the outside solar radiation has been
adopted. Leaf area index has been estimated weekly. Leaf area is composed by two terms,
the arched shoots part, considered as constant for the term of a month, and the flowering
shoots part. The arched shoots part has been measured monthly (4 plants samplings) and
(
)
[
]
VPDLAIGeE
LAIk
∗∗Β+∗−∗Α=
∗−
1
the flowering shoots part has been estimated on a weekly basis, as a linear function of the
length of the shoots, entered as input to the transpiration model. In the period considered,
LAI has varied between 2.12 and 2.51 and the flowering shoots part has represented from
22 % to 27 % of the total LAI. The following relation for the leaf area of the flowering
shoots has been obtained:
LA shoots (cm
2
)= 22.272 * L (cm) – 468.548 (R
2
= 0.85, n=67),
L= flowering shoots length
It has also been found that only 11 shoots are needed to estimate the leaf area of
the flowering shoots, with an error less than 5 %, but usually some more have been used.
The radiation extinction coefficient (k) has been considered constant throughout
the period and equal to 0.64 (Stanghellini, 1987).
Results and discussion
Transpiration Flux in Relation to the Microclimate
When transpiration (E) is considered in relation to values of outside solar radiation
(figure 1) between 100 and 400 wm
-2
, a positive correlation is observed, with lower
scattering than in the range of 400 to 600 wm
-2
. Estimation of E when solar radiation is
very low or cero is improved if the saturation deficit of air is considered, particularly
when required time intervals are short (Jolliet and Bailey, 1992; González, 1995). The
scattered values of transpiration for radiation values higher than 400 wm
-2
, could be
associated to stomatal conductance variations in response to VPD changes, depending on
fog system capacity.
Transpiration Model
The fitting of estimated and measured values of transpiration is shown in (figure
2) and is quite good all along the measured transpiration flux, which range is fairly wide,
between 20 and 230 g/m
2
.h (5 and 58 g/plt.h). Some degree of underestimation is
observed for the higher transpiration values, and an overestimation for the lower values,
though very acceptably fitted in the measured range. Similar or bigger deviations are
reported by Baille et al. (1994b) for a number of pot plants.
A very good prediction is obtained in the 24 h cycle when the course of the
measured crop transpiration values is compared to the estimated one for Autumn and
Winter (figure 3). The observed deviations betweeen the two sets of values, correspond to
model underestimations in the hours when the flux is maximum. Also an overestimation
is shown along the morning hours and, again, underestimation towards the sunset.
In general terms the prediction is considered good. The model works on a real
time basis, and performs with a high level of accuracy the reactions of the transpiration
flux to the climatic variations, solar radiation and VPD, during the day and the night. In
the night hours, when forced air heating equipment is not working during longer intervals,
due to milder temperatures, and measured crop transpiration flux is showing wider
oscillations, the model has been able to perform this evolution. These conditions can be
distinguished from the ones when aerotherms have been working more continuously, due
to lower temperature levels (figure 4). Also during the night there is a very good
behaviour of the estimated transpiration in relation to the course of VPD. Transpiration
values follow in real time the evolution of VPD, which oscillates depending on air forced
heating equipment on and offs (figure 4). A small anticipation of some minutes is
obtained for the model in relation to the measured crop transpiration. For instance, on a
sunny day, the quantity of supplied water has been about 1000 litres versus 2280 l
supplied by means of the time scheduling policy. This difference is much more important
when a cloudy day is considered, with only 520 l watering with the algorithm control.
Predictive Control of Water Supply
When the algorithm based on the studied transpiration model, has been
implemented and tested for the irrigation scheduling of the rose soilless crop on perlite, a
good performance has been obtained in the different conditions along Autumn and
Winter. Two different insolation days are shown (figure 5). Watering is supplied at
variable intervals, when the estimated transpiration integral reaches the set value. During
the night the crop transpiration rate and integral are estimated by the model, and the
required waterings are supplied to the crop. Measured crop transpiration all along the
night hours is significant, particularly due to force air heating, ranging from 24% in a
clear day to some 46% for an overcast day. Jolliet and Bailey (1992) report on some 22%
the night transpiration fraction of the 24 hours period, for young tomato plants, under
heated greenhouse, and Medrano (1999) obtained 10 to 20% for cucumber in Spain, in a
cold greenhouse in Autumn. Figure 5 shows a good fitting of the model in the short time
intervals, for two very different solar radiation days in Winter.
Water supply can be very well adapted, in this way, to actual plant requirements
all the time. Estimations made by the model have been very acceptable, at the practical
level, for irrigation control, in soilless growing systems that have a high level of fragility
due to the very low buffer capacity for water and minerals.
Acknowledgements
This work has been partially supported by projects 1FD97.0974.C0201 and
SC00.093 of the Ministry of Science and Technology and INIA. Contributions of J.Cerdá,
C.Arnal, A.Tomás, B.Hueso and A.Gallego are recognised.
References
Bailey, B. Montero J.I. Biel C. Wilkinson D.J. Anton A. Jolliet O. 1993. Transpiration of
Ficus Benjamina. Comparison of measurements with predictions of the Penman-
Monteith model and a simplified version. Ag. For. Met. 65: 229-243
Baille M. Baille A. Delmon D. 1994a. Microclimate and transpiration of greenhouse rose
crops. Agric. For. Meteo. 71: 83-97
Baille M. Baille A. Laury J.C. 1994b. A simplified model for predicting evapo-
transpiration rate of nine ornamental species vs. climate factors and leaf area. Sci.
Hort. 59:217-232.
Boulard T., Jemaa R. 1993. Greenhouse tomato crop transpiration model application to
irrigation control. Acta Hort. 335: 381-387.
González M.M. 1995. Estudio y modelización de intercambios gaseosos en cultivo de
rosas bajo invernadero. Tesis doctoral, Univ. Pol. Valencia, Spain, 192 pp.
Jolliet O., Bailey B.J. 1992. The effect of climate on tomato transpiration in greenhouse:
measurements and models comparison. Agr. For. Met., 58:43-63.
Lorenzo P., Medrano E., Sanchez M.C., 1998. Greenhouse crop transpiration. An
implement to soilless irrigation management. Acta Hort. 458: 113-119.
Medrano E. 1999. Gestión del riego en cultivo de pepino en sustrato: Evaluación de la
transpiración durante la ontogenia. Tesis Doc. Univ. Polit. Madrid. 216 pp.
Medrano E., Lorenzo P., Sanchez M.C. 2001. Evaluation of a greenhouse crop
transpiration model with cucumber under high radiation conditions. Acta Hort. 559:
465-470.