Optimal design of dynamic experiments in postharvest storage of produce

Optimal design of dynamic experiments in postharvest storage of produce

Arno Strouwen, KU Leuven

Fresh fruit and vegetables are perishable and need to be stored at appropriate conditions after harvest.

The temperature is typically set as low as possible but above the freezing point of the product, beyond which massive cell damage occurs and the product no longer can be considered as fresh.Low temperature storage reduces respiration and many biochemical reactions that are associated with postharvest ripening or senescence.

A further reduction of the respiration rate is possible by reducing the O2 partial pressure of the storage atmosphere and increasing its CO2 partial pressure. This is typically achieved through active control of the composition of the storage atmosphere (Controlled Atmosphere (CA) storage). The optimal gas composition is critical, as too low an O2 partial pressure in combination with too high a CO2 partial pressure induces fermentation in the fruit. This causes off-flavors (e.g., ethanol) and storage disorders (e.g., flesh browning).

Alternatively, products can be stored in so-called Modified Atmosphere (MA) packages; by carefully selecting the packaging material in such a way that the O2 and CO2 permeability of the packaging material matches the O2 and CO2 consumption and production rate of the product, respectively, at equilibrium a storage atmosphere composition can be achieved that is close to the optimal one.

The optimal storage conditions (temperature, O2 and CO2 partial pressures) of fruit and vegetables depend on the species, cultivar, ripeness stage and many other factors and must be experimentally determined. This is typically achieved by storing the product at many different combinations of temperature, O2 and CO2 partial pressures, and monitoring the change of quality attributes during the storage period which can be as much as one year for apple and pear fruit.

The Flemish horticultural auctions are introducing novel horticultural products at an ever increasing speed to outpace the international competition, and outsource the experimental determination of optimal storage conditions for fresh fruit and vegetables to the Flanders Centre of Postharvest Technology (VCBT, Leuven), which is managed by the co-promoter of this application.

This involves a large investment of capital resources (CA units) as well as labor, as these experiments need to be repeated during several seasons.

Also, novel storage technologies such as Dynamic Controlled Atmosphere (DCA) are emerging; these techniques involve an O2 partial pressure setpoint that dynamically changes during the storage period depending on the physiological response of the fruit.

The traditional methodology for determining optimal storage conditions is insufficient to optimize such storage protocols.

As an alternative, a model based approach to optimize storage protocols is being developed by MeBioS division of KU Leuven in collaboration with the VCBT.

This approach relies on comprehensive mathematical models that describe the behavior of the product as a dynamical system with inputs (temperature, O2 and CO2 partial pressures) and outputs (respiration and fermentation rate, quality attributes).

A key feature of such a dynamic model is the respiration kinetics which is traditionally described by a nonlinear model of the Michaelis-Menten type. This model involves biochemical parameters that must be determined experimentally on an individual fruit basis.

The experimental effort has, therefore, shifted from large scale experiments using many samples of fruit and based on a classical response surface modeling approach, to small scale laboratory experiments involving enzyme kinetics and nonlinear parameter estimation.

To collect the data required to estimate the Michaelis-Menten models' parameters, experimental tests have to be performed in which the fruit tissue is exposed to a set of different combinations of O2 and CO2 partial pressures as well as temperatures. Typically, this is done according to a full factorial design with multiple but constant O2 and CO2 levels. After each experimental test, the O2 consumption and CO2 production rate are measured using an appropriate technique (gas chromatography, optical sensors), and the model is fitted to the data using a nonlinear parameter estimation technique, such as the Levenberg-Marquardt algorithm.

This non-dynamical experimental approach has three major disadvantages, limiting the efficiency of the current experimental procedures in fruit storage experiments: Because of the large number of O2, CO2 and temperature combinations, this procedure is tedious. The cost of the individual tests typically limits their number and, hence, the information content of the entire experiment; The various O2, CO2 and temperature combinations are tested consecutively using different fruit samples, which introduces a large amount of biological variation in the complete set of experimental data; Constant O2, CO2 and temperature levels do not excite the fruit's respiration as much as fluctuating ones.

To increase the experimental efficiency in fruit storage experiments, the respiration pathways can be excited dynamically.

In other words, rather than keeping the O2, CO2 and temperature conditions constant during an entire test, the O2, CO2 and temperature can be modified dynamically throughout any given test.

This should be done in a systematic fashion, according to a dynamic experimental design, such that the information content of the entire experiment is maximized.

This will then result in more precise estimates of the parameters of the Michaelis-Menten models, and therefore also in more precise models for the respiration and the fermentation of the product.

This approach has two major advantages:

- The first major advantage of a dynamic excitation of the tissue is that the full range of experimental O2, CO2 and temperature conditions can be implemented in a single experimental test. As a result, fewer experimental tests will be required, without any loss of information;

- The second advantage is that the biological variation is no longer confounded with the effects of the O2, CO2 and temperature conditions, since an entire dynamic test can be performed on a single fruit sample.

Optimal Design of Dynamic Experiments:Towards more efficient experimentation in postharvest storage of fruit and vegetables
Arno Strouwen, KU Leuven,

María Dolores OrtoláDeterminación de la tasa respiratoria de frutas