Results & Discussion
I. regional variances in growth parameters
The overall tree ring widths share similar pattern among provenances: trees from all 5 regions showed sharp decrements of tree ring widths in year 1999 and 2002(Fig 1). It suggests that certain disturbances happened in these two years which negatively influenced the growth of white spruce from various origins in the common garden(will be discussed later). However, trees from each region did exhibit differences in ring widths(Fig 1); generally speaking, white spruce origin from local Alberta have wider tree rings comparing with individuals from other regions; on the contrary, seeds from Nova Scotia and northwestern area have narrower rings indicating low vitality and growth rates. In year 1995, average ring width of trees original from Alberta is up to 4.8 mm, but individuals from Nova Scotia and Northwestern regions only reached nearly half of this number. Seeds came from Alberta also have better performance under disturbances, suggesting high resistance and resilience to extreme and less suited climatic conditions. Therefore, local seeds source is favorable to plant under local climatic conditions. By tree ring widths, annual basal area increment (BAI) can be computed for trees from each region(Fig 2). Same conclusions can be drawn that local provenances have the largest stem volumes annual increment.
Populations have significant variances in height, diameter at 1.3m (DBH), and mean basal area increment(BAI) among regions(Table 5). The numbers in Table 5 represent the mean value of each variable, and the letter behind each value suggest significant differences between group. The populations from Alberta and Ontario area had significantly higher larger diameters and basal area increment comparing with the rest. In contrast, those from Nova Scotia and Northwestern exhibited lower growth and smaller size.
II. population-specific growth/climate response
Correlation coefficients between ring width and climatic factors(temperature(Fig3. a) and precipitation(Fig3. b)) were calculated on species-level(ignore provenances). The result(Fig 3) indicates that tree ring width is negatively related with the temperature of previous year August, and also positively correlated with the precipitation of the same month. Both decreasing temperature and increasing precipitation will reduce water deficit, which suggests that summer water accessibility is one important limiting factor for white spruce growth. Precipitation in previous October also show statistical significant to ring growth(Fig 3. b). However, tree growth is nearly stopped in winter time and how the precipitation in previous year October influence current year tree ring growth is still unknown; Since this study only have 5 samples per provenances, small samples size may induce false significant.
Then ring width-climatic correlation were also calculated among 5 regions. According the result(Fig 4) , white spruce from all regions share negative correlation with previous August temperature, and the range of coefficients of correlations is from 0.4-0.5. The differences of ring width-temperature relationships among regions are not obvious than I expected.
However, the response of ring widths to precipitation tells a different story. Origin-specific ring growth-precipitation relationships exist(Fig 5). The growth of Individuals came from Alberta have statistically significant correlation with precipitation of previous year August; meanwhile, trees origin from Nova Scotia showed negative relationship with October precipitation; seeds from Ontario regions have both positive relationship with previous summer precipitation and negative correlation with previous October. The various growth-precipitation responses can be linked with the climatic environment of the provenance areas. Since there are various growth-climatic relationships among provenances/regions, provenance-specific models(Fig 6) can be build to predict the performances of white spruce from certain provenances planted under certain climatic conditions.
So far, the results match the hypotheses that regional variances exhibit in growth and climate responses. The variances performances among populations (under the identical growing environment) are caused by genetics differences which are evolved to fit the climatic conditions of their origins. Therefore, I built model between population growth performance(in this case, BAI is used) and their source climate conditions(mean annual temperature, mean summer precipitation, summer heat moisture index and climatic moisture deficit)(Fig 7), to identify the source climate values for the most 'suitable' provenances for Calling Lake and its surrounding regions with similar climate condition. The colors of points represent the regions of sources; the error bar is the standard error, and the gray polygon is the 95% confident interval. Based the fitted models, the points with maximum BAI were highlight by red dotted line. Mean summer precipitation of seed sources has highest fitness to BAI(Fig 7.B), and r2 equals to 0.86; the rest three models can also explained at least roughly 50% of the variances of BAI among populations. According to models, the 'best' seed sources for Calling Lake should be: 2.2 °C mean annual temperature, 339 mm mean summer precipitation, the value of summer heat moisture index and climatic moisture deficit close to 51 and 148, respectively.
III. drought tolerant indices
According Fig 1, tree ring widths of all 5 provenances sharply reduced in year 1999 and 2002, but what happened in these two years? To find the possible explanations, I analyzed daily climate data of Calling Lake from 1961 to 2007. Comparing with normal climate(1961-1990 climate is assumed as normal and able to represent the long-term local climate pattern, the 0 lines in Fig 8), there are more extreme cold days(the blue area under the 0 line in Fig 8.A) than expected especially in the spring of 2002. Cold spring is very likely to cause frozen damage and decrease the length of growing season, and consequently decreases tree ring width of the year. Besides short growing season, water limitation is the other reason for sharp reducing tree ring widths in 2002. Fig 8.B and C showed low precipitation and high moisture deficit in that year.
However, the daily temperature of 1999(late 1998 to late 1999) is normal comparing with the local long-term temperature record. Since biologically temperature and precipitation are the two most important factors influencing plant growth, I compared the precipitation of 1999 to long period local precipitation data(Fig 8.B). Low precipitation in 1999 summer time cause the high moisture deficit, which can directly limit tree growth.
To summarize, trees from all regions had dramatic decrease of ring widths in year 1999 and 2002 with different reasons; one(1999) is for drought and the other year(2002) is caused by low temperature and drought(Fig 8).
To summarize, trees from all regions had dramatic decrease of ring widths in year 1999 and 2002 with different reasons; one(1999) is for drought and the other year(2002) is caused by low temperature and drought(Fig 8).
To exterminate the regional-specific responses to drought events, raw tree ring widths data were standardized and detrended to eliminate the age-effect and to highlight trees' response to the disturbances. The detrended ring widths of 5 regions(Fig 9) suggested various responses to disturbances among regions. For example, Ontario samples seems had severer disturbance in year 1999 than 2002; and those from Northwestern had very similar performances in these two years. The other interesting point is that Alberta local white spruce had decreasing ring widths after disturbances.
Using detrended ring widths, drought tolerant indices were computed for both year 1999 and 2002(Fig 10). Generally speaking, trees suffered more in year 1999 than 2002, but the difference among regions are mild. Differences responses in those two years may caused by lower severity disturbances in year 2002, or winter spruces are more sensitive to drought damage than cold. There are multiple other possible explanations for this result, and further study is needed.