Project
title / Environmental optimisation and the predictable production of
protected edible crops / DEFRA
project code / HH1329SPC

Department for Environment, Food and Rural Affairs CSG 15

Research and Development

Final Project Report

(Not to be used for LINK projects)

Two hard copies of this form should be returned to:
Research Policy and International Division, Final Reports Unit
DEFRA, Area 301
Cromwell House, Dean Stanley Street, London, SW1P 3JH.
An electronic version should be e-mailed to
Project title / Environmental optimisation and the predictable production of
protected edible crops
DEFRA project code / HH1329SPC
Contractor organisation and location / Horticulture Research International
Wellesbourne
Warwick, CV35 9EF
Total DEFRA project costs / £ 367,758
Project start date / 01/10/00 / Project end date / 30/09/03
Executive summary (maximum 2 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.

CSG 15 (Rev. 6/02) 2

Project
title / Environmental optimisation and the predictable production of
protected edible crops / DEFRA
project code / HH1329SPC

Supermarkets require reliable supplies of high quality fruit in agreed quantities. At present, however, growers lack the ability to be able to predict their tomato yields with a high degree of accuracy, or to manipulate their pattern of yield to a pre-determined schedule. Consequently this research was carried our to aid the development of systems to predict yields and to provide information to help manipulate the pattern of crop yields.

Modern glasshouse control computers enable heating and venting set-points to be changed dynamically to control humidity, or with intercepted light integral with a view to optimising the aerial environment and saving energy. However, it was unclear whether such control algorithms could be used to stabilise mean diurnal temperatures and, therefore, help control the pattern of yield. To address this issue an experiment was carried out where night temperatures were modified according to the amount of solar radiation intercepted the previous day. Night temperature were either set to 16°C and increased linearly by 3°C for every 1000 klx h intercepted outside during the previous day or set to 19°C and decreased by 5°C for every 1000 klx h. While the second regime might have been expected to help stabilise the mean diurnal temperature, it still fluctuated over the course of the experiment. This was because in summer it was not possible to reduce the night temperatures enough to compensate for warm sunny days. The relatively long days and short nights added to this problem. As the environments were similar in these two treatments so was the pattern of yield, there were only a few weeks when yields were significantly different.

Removing leaves from the canopy may increase fruit temperatures by exposing fruits to more solar radiation, thereby, affecting fruit ripening and the pattern of yield. Therefore two experiments were carried out. The first investigated the effects of removing different amounts of old leaves at the bottom of the canopy. Removing more leaves increased the temperature of older fruits during the day, although there was a tendency for these fruits to be cooler at night. The net result was that fruits at the bottom of the canopy were on average 0.2°C warmer. This had the effect of hastening fruit ripening by 1.2 days. However, this treatment only had a small effect on the pattern of yield. Surprisingly, even though the high leaf removal treatment had around half the leaf area (12 fewer leaves) of the low leaf removal treatment, there was no significant loss of yield over the course of the season. This was because lower leaves were contributing little to net canopy photosynthesis due to that fact they were intercepting very little light. However, increased leaf removal reduced the water uptake of plants. A second experiment investigated the effect of removing one in every three young leaves. Again this treatment had little effect on the pattern of yield, although reducing the leaf area by ~28% resulted in an 8% loss of yield. This was because light which would have been intercepted by young efficient leaves at the top of the canopy was instead intercepted by older leaves which were around 25% less efficient.

Models of fruit development/ripening time are needed to enable accurate yield prediction models. Earlier work indicated that the sensitivity of fruits to temperature interacted with their stage of development. In particular late stages of development appeared more sensitive to temperature. However, more data was needed to adequately quantify this response. Therefore, two experiments were conducted. In the first, growth rooms containing tomato plants were changed from 20°C to either 15 or 25°C. Fruits that were harvested before the temperature regimes were changed ripened on average after 56.3 days. Fruits exposed to 15°C showed a progressive delay in fruit development time as the duration at 15°C increased. Similarly the longer fruits spent at 25°C the quicker they ripened. Once fruits that had flowered at either 15°C or 25°C were being picked fruit development times remained fairly stable; on average fruits ripened after 91.6 and 44.3 days at 15°C and 25°C, respectively. In a second experiment the effect of a pulse of high or low temperature on fruit development times and the pattern of yield was examined. Due to the fact that the treatments were only given for a relatively short duration (1 week) more extreme temperature regimes were chosen (10°C or 30°C). When fruits were grown continuously at 20°C they ripened on average after 57.3 days. Around 5 days after the beginning of the temperature pulse the fruit development times began to diverge. The fruits exposed to 30°C showed a hastening and then gradual delay in fruit ripening times. The opposite was true for fruits exposed to one week at 15°C. An increase in yield was observed following the high temperature treatment, however, this was more than compensated for due to lower yields in future weeks. Rather than an increased sensitivity to temperature with the stage of fruit development, these data indicated a change in the optimum temperature with stage of fruit development. Temperatures above 20°C delayed the development of very immature fruits, but hastened the development of more mature fruits.

A model was developed to predict crop yields. The fruit development model was used to predict the pattern of yield based up on when fruits were estimated to ripen for any given set of daily glasshouse temperatures. Yields were then estimated using recorded fruits loads, shoot densities and mean fruit sizes. The model was tested using two years of historical data. The model gave reasonable predictions of the pattern of yields and accounted for 75% and 90% of the variance in weekly yields. These simulations with historic data used the recorded mean diurnal temperatures. However, when the model is used to forecast weekly yields in real time, forecast temperatures will be needed. Simulations showed the predictions were highly sensitive to the temperature data. Therefore, to aid decision making the model predicts the weekly yields based upon the estimated glasshouse temperatures and also what might happen should the weather change and higher or lower temperatures occur.

Modifying temperature set points based upon the light integral and leaf removal strategies had little effect on the pattern of yield. Therefore, the final experiment attempted to stabilise the yield by modifying the temperature set-points. Decisions were made having run simulations with the model and studied weather forecasts. Generally, set-points were lowered when high temperatures were forecast and raised when low yields were forecast. However, consistently achieving the target temperatures in a glasshouse proved difficult. Periods of high temperature could not be avoided which increased and subsequently decreased yields. In many of these cases the model predicted a decrease in yield could not be avoided unless extremely high temperature regimes were used. However, these set-points were not used due to fears that it would have a negative impact on the crop. Instead smaller more acceptable changes in temperature were used, however, these changes proved to be insufficient to have a significant effect on the pattern of yield. If hot sunny weather is forecast it may be possible to raise temperature set-points in advance of the sunny weather so as to hasten ripening and increase yields before/during the sunny weather when demand for salads is high, this would also help to reduce the magnitude of the peak in yield that would naturally occur after the hot weather. However, such a strategy is limited by the need for accurate weather forecasting for around two weeks ahead.

CSG 15 (Rev. 6/02) 2

Project
title / Environmental optimisation and the predictable production of
protected edible crops / DEFRA
project code / HH1329SPC
Scientific report (maximum 20 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.

CSG 15 (Rev. 6/02) 2

Project
title / Environmental optimisation and the predictable production of
protected edible crops / DEFRA
project code / HH1329SPC

1) Introduction and objectives

Supermarkets require reliable supplies of fruit in agreed quantities and of a standard quality. At present, however, growers lack the ability to be able to predict their tomato yields with a high degree of accuracy, or to manipulate their pattern of yield to a pre-determined schedule. Consequently this research was carried our to aid the development of systems to predict yields and to provide information to help manipulate the pattern of crop yields. Information of this type should help British growers to maintain supplies to the UK multiples.

Previous work (HH1326SPC) successfully provided an understanding as to why long season tomato crop yields fluctuate from week to week. While the pattern of fruit set and light integral affect the pattern of yield, ‘cyclical fruit production’ was primarily due to changes in the time fruits take to ripen. The effect of temperature on the time of fruit maturation was examined. A period of elevated temperature hastened fruit ripening and resulted in a flush of ripe fruits occurring shortly afterwards, although as a result there was a decrease in the number of fruits nearing maturity on the plant, resulting in depressed yields in subsequent weeks. This effect was largely due to the fact that fruits become more sensitive to increased temperature as they approached maturity. However, there was insufficient data to fully quantify this response, which proved to limit the accuracy of yield predictions given by a model that was developed.

One of our aims was therefore to investigate in greater detail the change in sensitivity of fruits to temperature and incorporated this information into a model (objective 3). The potential for using this model for commercial predictions would be explored by validating it and testing the dependence of the output on the accuracy of weather forecasts (objectives 4 and 5). While the accuracy of weather forecasting may limit the potential use of the model for predictive purposes, that does not preclude its use to indicate how growers might manipulate the glasshouse environment to control the pattern of yield.

Growers often increase their glasshouse temperature set-points in periods of high irradiance, and this practice may contribute to the problem of cyclical fruit production. Hence we also investigated whether it is possible to smooth out some of the fluctuations in yield by creating a more stable glasshouse temperature. We assessed the degree to which it is possible to reduce night temperatures following a warm sunny day (objective 1) and the effect this has on the pattern of yield (objective 2). Furthermore, the effect of cultural practices, such as deleafing were explored (objective 2) as the degree of shading may be important in determining fruit temperatures and hence when fruits ripen.

1.1. SCIENTIFIC OBJECTIVES:

1. To establish the feasibility of reducing night temperature set points following warm days so as to minimise the day-to-day variation in mean air temperature.

2. To quantify the degree to which it is possible to minimise weekly fluctuations in yield by manipulating the glasshouse environment and modifying the deleafing

3. To determine the sensitivity of fruits to temperature at different developmental stages, develop an improved system for predicting the time of fruit maturation and incorporate this into the current tomato model.

4. To validate the modified tomato model using data collected as part of objective 2, together with historic yield data and to assess how dependent the yield predictions will be on the accuracy of meteorological forecasts.

5. To use the model and understanding gained to attempt to manipulate glasshouse yields according to a pre-determined pattern.

All of the objectives were met in full and on time. A brief overview of the work conducted for each of these objectives follows.


2) To establish the feasibility of reducing night temperature set points following warm days so as to minimise the day-to-day variation in mean air temperature (Objective 1).

Although the air temperature within glasshouses can be controlled reasonably well through the use of heating and natural ventilation, cooling during hot weather is not yet cost effective. Consequently, during the summer, air temperatures will fluctuate from week to week causing fluctuations in the pattern of yield. However, modern glasshouse control computers enable heating and venting set-points to be changed dynamically to control vapour pressure deficit (humidity), or with intercepted light integral with a view to optimising the aerial environment and saving energy. Controllers are increasingly exploiting the fact that plants grown in a fluctuating temperature regime often suffer no overall loss of yield relative to those grown in a constant regime having the same mean temperature (Hurd and Graves, 1984; Khayat et al., 1985; de Koning, 1988, 1990). These systems make better use of solar energy to heat glasshouses and can enable fuel to be burnt when the energy losses are lower. However, it is unclear how modifying temperature set-points based on daily light integrals will affect the pattern of yield or whether such control algorithms can be used to stabilise mean diurnal temperatures and, therefore, help control the pattern of yield. To address this issue an experiment was carried out where night temperatures were modified according to the amount of solar radiation intercepted the previous day.