Seed production in Pinus radiata D. Don : Impact of climate and site on numbers of emergent female strobili.
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
This thesis describes a study undertaken to improve the capacity of seed orchard managers to select productive sites for the production of Pinus radiata seed. It is applied in its approach, with the major intent being to explore the relationship between environmental variables, and numbers of emergent female strobili at anthesis. Following the development of a sampling methodology for counting strobili, data were collected from 52 open grown stands of Pinus radiata ranging in age from 10 to 12 years located over a range of latitudes and altitudes east of the main dividing range of the South Island. Stands were selected to be reasonably close to a meteorological station. Based on existing knowledge of when environmental variables impact on the numbers of strobili carried at the time of anthesis, monthly temperature and water stress data for the spring months of the year prior to anthesis, and also for subsequent summer months, were correlated with numbers of strobili across sites. Winter temperatures prior to anthesis were also correlated with the strobilus data. The variables that were significantly correlated with strobilus number were then included in a stepwise multiple regression analysis with the intent of developing a statistical model for strobilus production. The multiple regression analysis selected the following variables into the model equation (with the percentage of the variance accounted for by each variable shown in brackets): (i) Mean minimum temperature for February (around the time of seed-cone bud differentiation) (54%). (ii) Integrated soil water stress for February and March (6%). (iii) Accumulated growing degree days weighted against rainfall for the spring prior to seed-cone bud differentiation (4%). (iv) Mean minimum temperature for May (between differentiation and anthesis) (1%). Overall the model was able to account for around 65% of between-site variability in strobilus number. The model variables suggested that predominantly temperature and to some extent rainfall govern the number of seed-cone buds that begin differentiating; and the number that finally reach antl1esis in the spring. Two years later in 1994 the statistical model was validated across 12 new South Island sites. It was found that the model estimates were in close agreement with actual counts, with 75% of the residual values falling within the 95% confidence limit bounds. Considering that mean February temperature was so highly correlated with strobilus number it was considered appropriate, whilst validating the model in the South Island, to extend sampling into the North Island. 11 stands were sampled, situated along both east and west coasts, where warmer summer temperatures prevail. However, the model estimates were less accurate in the North Island than the South. Both the data for 1992 and 1994 (total 74 sites) were combined and strobilus data were remodelled. Again the multiple regression analysis selected the mean minimum temperature of February as the variable that explained most of the between-site variance. The mean number of strobili at sites in the North Island was 1.4 times greater than that for the sites in the South Island, with trees at the best North Island site carrying 38% more strobili than those at the best site in the South Island. Considering that long-term data for the February temperature variable were readily available and that the relationship between temperature and strobilus production is likely to have an upper-limit, the Schumacher sigmoid growth function was fitted to the strobilus data for the 74 sites in the both the 1992 and 1994 studies. This function was found to satisfactorily describe the relationship between strobilus number and February temperature, although due to the lack of data points at the high end of the temperature range, the response curve was not typically sigmoid with a flattening off at higher temperatures. Using both a GIS mapping system and a climate surface model, mean minimum February temperatures for New Zealand were mapped to assist with the selection of potential seed orchard sites. It was apparent from the maps that prime sites for strobilus production are likely to be those that have a mean minimum February temperature >15°C, and that those sites tend to be on northern coastal headlands and capes with insular climates. The majority of the sites meeting the temperature criteria are in the North Auckland region. It is likely that a reproductive environment for Pinus radiata is one where the temperature does not vary greatly, either diurnally or seasonally. Assessments of conelet abortion and cone seed yield and weight, made at the time of counting strobili, showed that the levels of conelet abortion were independent of the numbers of strobili carried by trees. Seed weight was also unrelated to strobilus production, but seed number was correlated. The high strobilus-producing sites tended to yield more full seeds per cone. To verify the significance of the variables in the statistical model experimental studies were undertaken using clonal graft material at Amberley seed orchard, Canterbury. Firstly, an experimental study confirmed the importance in the model of February as a time to influence strobilus numbers. Grafts put into a low-intensity-light building for 2 week periods during February and early March produced no strobili the following spring, while strobilus production was unaffected on grafts moved at other times between January and the 26th of March. In a further experiment, cooling grafts during the summer months reduced the number of strobili per whorl by 44% and the number of grafts carrying strobili by up to 80%. Siting grafts at cool inland sites during the winter on the Canterbury plain caused a loss of up to 61% in seed-cone buds. However, neither warming grafts in a polyethylene house during long shoot primordial initiation/differentiation; or in a heated glasshouse during the winter months leading up to anthesis, increased strobilus production. The reasons for this are discussed. In hindsight it would appear that the maximum temperatures in the polyethylene house were too high for strobilus production, whilst in the heated glasshouse there was insufficient diurnal temperature variation, along with an interruption to the normal seasonal cooling pattern. Nonetheless, considering both the site predictions of the strobilus model and findings of these experiments it is proposed that an optimal reproductive environment for radiata pine is probably one where warm temperatures are relatively constant and do not vary greatly diurnally or seasonally. Summer soil water stress was implicated with strobilus production, although to a lesser extent compared to mean minimum February temperature. That is, the sites more water stressed during February and March tended to produce more strobili than those that were less water stressed. This effect was borne out by an irrigation pot trial in 1993 which showed that irrigation reduced the numbers of strobili carried by grafts by up to 45%. However, due to persistent summer rain this effect was not satisfactorily demonstrated in an irrigation field trial conducted at Amberley seed orchard. The temperature and moisture conditions during the spring of the year prior to anthesis were found to indirectly impact on strobilus production by modifying the number of branches produced by grafts. When graft shoot growth was increased by modifying spring growing conditions, more branches developed per whorl. Finally, the effect of nitrogen fertiliser applications on strobilus production was tested across a range of sites and climates. There was no significant direct effect of N-fertiliser on strobilus production, regardless of N-form applied. It is unlikely that the use of nitrogen fertilisers in a well managed radiata pine seed orchard will enhance strobilus production; or that the inclusion of a soil-N variable in the model would improve its predictability. The statistical model developed in this thesis has added to knowledge regarding the identification and the importance of site environmental variables associated with female strobilus production in Pinus radiata. Further the model identified mean minimum February temperature as a key variable, for which long term data is readily available across New Zealand. These data were used to construct 2 mean minimum February temperature maps for New Zealand which can now be used to assist with seed orchard site screening. Experimentally it was demonstrated that temperature and moisture conditions govern the number of seed-cone buds that are differentiated and reach anthesis. It is recognised that environmental variables can impact on strobilus formation in both direct and indirect ways. The possibilities for extending the databases developed in this thesis to develop further understanding of seed production processes is advocated.