A postal survey of 1060 growers was conducted in March 1990 in an attempt to identify management factors and features within orchards which might be implicated in postharvest fruit loss. The survey, conducted by staff from MAF, DSIR and the KMB (Catherine Richardson and Prue Burt) was part of a wider programme seeking to link the impact of various practices at different stages in the orchard-packhouse-coolstore chain, with national fruit loss records compiled by the KMB.

As expected, results of the survey indicated that there were no simple, direct 'cause and effect' relationships between management factors and fruit losses postharvest. Although a number of associations were observed, these were not consistent from one year to the next, and were not consistent between regions. Associations tended to be related to the predominant category of fruit loss in each season (eg., botrytis in 1988/89, and soft fruit in 1989/90).

Fruit loss

Orchardists responding to the survey accounted for 15% of the registered growers in 1989, 15% of the 15,590 ha of kiwifruit under cultivation, and produced approximately 20% of the export crop. Of the 7.7 and 11.1 million trays condition checked from these orchards in 1989 and 1990, botrytis (3.4%) and soft fruit (9.8%) were the primary categories of fruit loss. Losses in the sampled orchards were typical of trends in the national fruit loss data each season.

Management practices

Responses to the survey provided an intriguing overview of how New Zealand kiwifruit orchards are managed and highlight the vast range of possible combinations of practices used to grow this crop (Table 1). In fact, it was calculated that no more than three growers in the entire industry manage their crop in a similar way! Such diversity suggests that there is no such thing as the 'average' grower and also complicates analysis of how these practices influence fruit losses.

The way in which individual orchards were managed was very similar between the two seasons. The main differences were the increased use of budbreak enhancing sprays (Table 1) and artificial pollination (from 12 to 21%, with a corresponding decline in the proportion of growers using bees alone).

Associations between management practices and fruit losses

When orchard practices were matched against annual and regional fruit losses, they could be divided into two groups; one containing factors which were unrelated to fruit loss; and another where there were statistically 'significant' associations with some category of fruit loss (Table 2). While the list of orchard factors apparently related to fruit loss seems extensive, the associations were frequently inconsistent from one year to the next, and those in the Bay of Plenty often differed from those in other regions.

For example, in 1989, botrytis losses on orchards containing predominantly pergola support structures were greater than those with mainly T-bars (Table 3). In the subsequent season, however, it was soft fruit losses which were associated with the type of trellis structure, not botrytis loss. Many other factors listed in the left-hand side of Table 2 were also associated with the major category of fruit loss each season, ie, botrytis in 1989 and soft fruit in 1990. It is not surprising that such relationships are found only when large numbers of observations are involved.

Further inconsistencies were noted when orchards with different types of irrigation system were compared (Table 3). In 1989, botrytis losses were greater on orchards with drippers than those with sprinklers. The following year, losses on orchards with sprinklers were greater than those with drippers! Examples where trends reversed from year to year were less common than those where the association changed according to the predominant category of fruit loss.

In some cases, where factors could have a range of values (eg, fertiliser inputs, orchard production, orchard area), there were distinct trends with the fruit loss data. These are indicated by a (+) or (-) in Table 2. Botrytis losses, for example, increased (+) with vine age in 1989. The proportion of the property sprayed with budbreak enhancing spray, and the amounts of nitrogen or potassium applied, were each associated with botrytis losses in 1989, and with soft fruit losses in 1990. However, these relationships were not strong enough to be used to predict fruit losses, because of the amount of scatter in the data. Using the example of potassium and soft fruit losses in 1990 (Fig. 1), individual orchards with a potassium input of 150 kg/ha lost between 0 and 60% soft fruit! Individual factors such as potassium application accounted for no more than 4% of the variability in the data. In other words, 96% of the variation was caused by factors other than fertiliser. Relationships accounting for more than 80% of the variability would be required before they could be used reliably for predictive purposes within the kiwifruit industry.

Figure 1. Association between potassium application rates and soft fruit losses in 1990.

It is possible to put combinations of factors together to test whether they describe fruit loss better collectively than individually. Soft fruit losses in 1990 (Table 4) were analysed in three ways: to find factors describing losses in a data set comprising information from all the KMB regions, to identify factors describing losses in the Bay of Plenty alone, and those describing losses in KMB regions other than the Bay of Plenty. The upper part of Table 4 indicates that combinations of individual factors accounted for no more than 25% of the variability in the loss data. Furthermore, there was no consistency between the factors best describing fruit loss in one region, and those in another. In the lower part of Table 4 an additional factor was added to the analysis, the individual KMB region in which fruit was grown. Comparison of the upper and lower parts of Table 4 indicates that, by itself, the region in which fruit was grown was a superior predictor of fruit loss, ie., it accounted for a greater amount of the variability in the data than combinations of large numbers of individual management factors. This suggests that climatic and environmental factors specific to the growing region can exert a larger influence on fruit loss, than superimposed management factors which growers can better control.

Summary

We have examined a large selection of orchard management practices that could have potentially been related to postharvest fruit losses. A number of factors were found to be unrelated to fruit losses. Other factors were occasionally related to losses, but the relationships were absent, or even reversed, in the subsequent season. Even factors which have been shown to influence storage quality characteristics, such as waterlogging, failed to give reproducible associations in this survey.

Lack of conclusiveness and 'flip-flopping' between seasons may be partly due to the imprecise way in which the industry assesses and records fruit loss, and also the subjective nature of responses required by certain survey questions. These limitations should be offset to some extent by the large number of records in the database. Complete resolution as to whether specific factors are indeed related to fruit loss must wait for more careful experimentation in the future.

There were no indications from this study that changing current orchard practices might lead to significant improvements in the level of fruit losses. The observation that the region fruit is grown in seems to have a larger influence on losses than on-orchard factors, suggests that the influence of climatic and environmental factors on fruit development deserves further attention. In addition, a large discrepancy in losses can occur when fruit from the same block are packed on the same day and stored in different coolstores. Thus, while orchard practices have the potential to influence the storage characteristics of fruit, they are not the only culprit responsible for postharvest fruit losses.

Table 1. Standard practices in kiwifruit orchards in the 1989/90 growing season (Numbers indicate the percent of growers*)

Orchard Management
Predominant type of support structure:	Pergola	T-Bar	Other
	55	40	5
Predominant internal shelter species:	Willow	Casuarina	Cryptomaria	Poplar	Artificial
	59	32	29	19	9
Method of disposal of shelter prunings:	Mulched	Removed from orchard	Shelter not pruned	No natural shelter
	80	17	2	1
Method of weed control around vines:	Chemical herbicides	Mulching	Manual removal	Animals
	87	9	2	2
Type of irrigation system:	Drippers	Sprinklers	No system
	47	31	21
Time of initial or basal fertiliser application:	Dec-June	July-Aug	Sept-Nov
	18	74	8
Fertiliser application	Fertiliser applied	No N applied	No P applied	No K applied	No fertiliser of any sort applied
	95	18	15	9	5
Vine Management
Disposal of vine prunings:	Mulched orchard	Removed from herbicide strip	Left to rot in
	97	2.5	0.5
Budbreak enhancing sprays:	Proportion using spray	Proportion of property sprayed
1988	14	5%
1989	45	30%
Method of pollination:	Bees only	Bees and artificial pollination	Artificial pollination only
	76	21	3
On-orchard males	M-series	Matua	Tomuri	Other
	51	57	11	8
Distribution of male vines:	Interspersed	Overhead	Strip
	70	17	13
Method of summer pruning:	Manual	Mechanical	Summer pruning omitted
	86	7	7
Dormant cleanup sprays:	No sprays used	Copper based sprays	Lime sulphur	Both copper and lime sulphur
	41	39	11	4
Harvesting Operations
Method of harvesting:	Contract pickers	Staff (and family) employed by grower	Family members only
	63	30	7
Location of packhouse in relation to orchard:	Packhouse on-site	Packhouse <5km from orchard	Packhouse 5-50 km from orchard	Packhouse >50 km from orchard
	14	39	43	4
Location of coolstore in relation to packhouse:	adjacent to packhouse	<5 km from packhouse	5-50 km from packhouse	>50 km from packhouse
	48	14	33	6
* Based on a sample of 626 growers.

Table 2. Factors in the orchard environment which may influence postharvest fruit losses.

Factors with 'significant' associations with some category of fruit loss	Factors not associated with fruit loss.
* Natural shelter species	* Shelter (artificial vs natural)
* Production (+) (export trays/ha)*	* Method of disposal of shelter prunings
* Timing of fertiliser applications (-)	* Irrigation (no system vs system present)
* Amount of nitrogen or potassium applied (+)	* Fertiliser (fertiliser applied vs no application)
* Orchard area (+)	* Amount of phosphorus applied
* Vine age (+)	* Method of disposal of vine prunings
* Type of support structure	* Pollen source (commercial vs on-orchard males)
* Type of irrigation system	* Distribution of pollinators
* Method of weed control	* Summer pruning (Pruning done vs pruning omitted)
* Proportion of property sprayed with budbreak enhancing sprays(+)	* Number of leaves left after last fruit when summer pruning
* Method of pollination	* Presence of pests
* Species of on-orchard pollinator	* Dormant cleanup sprays (sprays applied vs omitted)
* Factors affecting vine performance - nutrient stresses	* Factors affecting vine performance - waterlogging, drought, frost damage.
* Presence of diseases	* Source of staff employed to harvest fruit
* Type of dormant cleanup spray	* % rejection rate of fruit at packhouse
* Time orchard usually reaches maturity
* Location of packhouse relative to orchard
* Location of coolstore relative to packhouse

Table 3. Influence of trellis structure and irrigation system on fruit lost during repacking (% of fruit condition checked between August and November in 1989, and 1990).

	Botrytis			Soft
Year	Pergola	T-Bar	aP	Pergola	T-bar	P
1989	4.1	2.4	***	2.1	1.4	ns
1990	1.2	0.9	ns	11.7	7.3	***
	Dripper	Sprinkler		Dripper	Sprinkler
1989	3.8	3.0	*	2.0	1.7	ns
1990	0.9	1.5	*	11.4	8.0	*
aP = probability; *** = a 99.9% , or * a 95% chance that the two numbers to the left being compared are different; ns = no significant difference between the two numbers to the left of the symbols.

Table 4. The most promising factors for predicting soft fruit losses in 1990.

			% of data variability accounted for+
A. KMB region excluded as a factor
1. All regions	*Production (1990)		3.9
	Time of fertiliser application		7.0
	Mean fruit size		9.1
	Method of disposal of shelter prunings		10.6
	Potassium application (kg/ha)		12.0
2. Bay of Plenty	Production (1990)		8.8
3. All other regions	Use of budbreak enhancing spray		15.1
	Production (1990)		19.0
	Potassium application (kg/ha)		22.1
B. KMB region included as a factor
1. All regions	KMB region		13.2
	Production (1989)		17.5
	Method of shelter pruning disposal		18.9
2. Bay of Plenty	Production (1990)		8.8
3. All other regions	KMB region		35.4
	Production (1990)		38.9
* Production measured in export tray equivalents per hectare. + Note, factors are added successively until there is no further improvement in the amount of variability accounted for. The most important factor is that which is first in a list, and the inclusion of each additional factor contributes to an increase in the amount of variability accounted for. Predictive 'models' that are potentially useful account for more than 80% of the variability.