Lincoln University >
Research Archive >
Theses and Dissertations >
Doctoral (PhD) Theses >
Cite or link to this item using this URL:
|Title: ||Apple tree system research|
|Author: ||Zhang, Jianlu|
|Degree: ||Doctor of Philosophy|
|Institution: ||Lincoln University|
|Item Type: ||Thesis|
|Abstract: ||This has been a co-operative research project between the Ministry of Agriculture and Fisheries, Division of Horticulture and Processing, DSIR, New Zealand Apple and Pear Marketing Board, The New Zealand Fruit growers Federation and the Department of Horticulture, Lincoln University commenced in the winter of 1986. The objective was to monitor 15 Royal Gala or Gala apple orchards in the three main apple producing regions of New Zealand to produce a data base for modelling biological and financial interactions. Selected trees were continuously monitored for three growing seasons. One orchard was continuously monitored for another 3 years (in the third additional year for the purpose of model validation). The orchards were 5 - 8 years old during the first monitoring season. Twelve of the orchards were on MM106 rootstock and the remainder on M793. Tree density ranged from 455 - 1102/ha. Five trees were monitored for fruit number and fruit quality in each orchard at harvest. Three branches from each monitored tree were chosen to record the number of flowers and fruits before thinning and after thinning.
To establish a fruit tree branch sampling system which can be reliably extrapolated to a whole tree basis, measurements from 204 trees, 151 parent branches and 9283 sub-branches were involved. Wood density was measured and the estimation of the volume of a branch, and the ratio of the branch to the whole tree, compared. Evidence is presented to show that the ratio of the sum of branch cross sectional area (CSA) to the trunk CSA is equal to the ratio of fruitfulness of the whole tree to fruitfulness of the sum of the branches. A relationship has been established between the sum of the CSA of branches directly arising from the central leader and the trunk CSA of central leader-trained apple trees. Ratios of 1.6 - 2.1 : 1 have been determined over 7 years of research without any major influence from pruning. The relationship of secondary branching of individual fruiting arms was investigated also, suggesting a similar relationship between branch CSA and secondary branch CSA but the influence from pruning, in this case, is relatively greater. Between orchard, tree and branch variation is explained. Recommendations on branch sampling for research and monitoring purposes are provided, based on these findings.
Climatic data from 7 meterological stations was used, incorporating temperature, sunshine hours and rainfall. For each season the orchards were monitored, climatic data was recorded for the period October in the previous growing season until harvest time (March) of the current season, a total of 18 months in each case. Correlation calculations were attempted for all monthly combinations. Contour maps of correlation coefficients between tree parameters and monthly combinations of climatic data are presented to show the major effects of climate on fruit production. The production of multiple regressions has been focused on a range of tree parameters, with relevant climatic data from the contour maps added, to refine the mathematical relationships. Higher maximum temperatures in the period of March and April produced higher initial fruit set (r = 0.54**) the following spring suggesting an effect on quality of the flower. Lower maximum temperatures in the late dormant period (August to September) produced a higher initial fruit set (r = -0.40**) as well. The natural simple regression between initial fruit set and flower number per cross sectional area (CSA) gives an r value of -0.47** indicating that the more flowers on a tree the lower the percentage initial fruit set. Introduction of the climatic effects produces a refined multiple regression with an r value at 0.76**. A negative correlation for estimating fruit number/ha after thinning was determined (r = -0.78**).
The estimation of fruit volume from fruit diameter was explored. Fruit growth curves based on fruit diameter or volume were plotted for 2 growing seasons. The influence of temperature on fruit growth is documented. Fruit size variation was explored and the sampling size for estimating average fruit weight is recommended. It is demonstrated how fruit number after thinning can be extrapolated to estimate total yield, fruit size and size distribution at harvest. Average fruit size was positively correlated with November to February temperatures in the previous season and also maximum temperatures in the December to January period prior to harvest. A negative correlation with fruit size was found for minimum temperatures in May. All these improved knowledge on the relationship between average fruit size and fruit number after thinning, but the maximum temperatures in December and January were critical in providing more accurate information (r = 0.81**) for harvest predictions.
For management planning it is necessary to relate yield, fruit quality and size distribution, revenue and cost data, in order to calculate annual gross margins. A set of dynamic mathematical models developed are presented, to demonstrate the interrelationship of factors influencing apple production and profitability. Based on the model, a computer program was produced for practical orchard application. At the blossom and pre-thinning stages of fruit production, model users may specify parameters recorded from their own trees (eg flower numbers) and climatic data to predict tree behaviour for the next stage. At the post-thinning stage, model users may again specify parameters recorded from their own trees (eg fruit number) and climatic data to predict yield and fruit size distribution at harvest. These predictions can be correlated with financial data such as price realisations for the fruit and production and harvesting costs, to minimise estimated net return. The models allow annual climatic data to be balanced against biological parameters (eg fruit number) in order to minimise costs and maximise gross income and annual gross margins. For example, the required fruit number per tree after thinning can be correlated with the size grade distribution which will maximise returns, incorporating weather, cost and other key management parameters.
Carbohydrate reserves in the dormant period are of primary importance for fruit production in the following year. The storage of starch in different parts or organs of apple trees was compared. The concentration of starch in the root system is much higher than that in the above ground parts. Observations of root growth and photos provided the evidence that roots continued to grow in the dormant period in Canterbury. Non-structural young rootlets contained almost no starch. For secondary roots, on average, thin roots contained higher concentrations of starch than thicker roots. However, results indicated there is greater variation of starch concentration in thinner roots. Results also showed that the starch content in root bark is much higher than that in root wood. The sampling variation may be reduced when bark and wood are separately tested. This is because the ratio of bark to wood is different for different roots. To minimize the variation, the bark of roots above 1 cm in diameter is recommended for analysis. Under 3 different crop loads, starch content was measured using replicates of 6 roots per tree and 3 trees per treatment. Starch content under light crop load conditions was significantly higher than that of middle and high crop loads, indicating a negative correlation with crop load in the growing season. A negative relationship between the content of starch and soluble sugar was found, reflecting the vigour of root growth. For the following year's production, higher yield or fruit number was positively correlated with the stored starch level or negatively correlated with the yield or fruit number in the previous growing season.
An accurate method of estimation of fruit size distribution using average fruit weight and standard deviation of the mean is discussed.|
|Supervisor: ||Thiele, G. F.|
Rowe, R. N.
|Persistent URL (URI): ||http://hdl.handle.net/10182/1878|
|Appears in Collections:||Doctoral (PhD) Theses|
Department of Agricultural Sciences
Copyright in individual works within the Research Archive belongs to their authors and/or publishers. You may make a print or digital copy of a work for your personal non-commercial use. Unless otherwise indicated, all other rights are reserved, except for other user rights granted by the copyright laws of your country.
If you believe that copyright is being infringed by material available in this archive, contact us and we will investigate.