Graphics and results

Clinal trend of Brown-down date and bud set

The map of provenance trials shows a longer growing season change of MN seed source as they set buds later than others. In the map below, higher budset score means an earlier stage while lower ones are later. The points are the sources of provenances, while the values are recorded in the Athabasca trials. Seed sources from northern British Columbia and Alberta have more red color, while MN shows a significant delay as the green color. MN source set buds later which may suffer a potential early frost risk in fall. The brown-down patterns agree with the bud set trend from north to south, which may indicate an adaptive trait of growth cessation before winter dormancy for both latitudinal and climate factors.

In Figure 1, the temporal change of brown-down from 2001 to 2006 was based on the EVI data. Warm-color points indicate earlier Julian dates, while green ones are late brown-down which is after  the date of 300 (the end of October or the beginning of November).  Green areas are more clustered in the Prairie Provinces and the Minnesota State, which indicates a late bud set date. Warmer color points in the north show an earlier bud set date. This north to south gradient is similar as the provenance trial records. However, the pattern per year still varies from each, which provides evidences of some factors other than latitude gradient.

Figure1 Bud set score of provenance trials and brown-down date (2001-2006)

Brown-down date versus longitude and latitude 

Figure 2 shows a strong clinal pattern with latitude or longitude among aspen populations.  The nothern populations (TP) have later brown-down date than the southern ones (BS).   The five color legends  refer to five eco-regions in the study area. 

Figure 2 Scatter plots of mean brown-down date versus longitude and latitude.   The miniature map shows the locations of different ecoregions. 

Bud set versus longitude, latitude and elevation

Figure 3 also shows a clinal pattern with latitude or longitude among 43 provenances.  The nothern provenances (TP) have later brown-down date than the southern ones (BS).   The five color legends  refer to five eco-regions in the study area as in Figure 2. 

Figure 3 Scatter plots of mean bud set score versus longitude and latitude.   The color legend is the same as in Figure 2. 

Annual brown-down date variation

For brown-down date, there is distinctiveness in different years (Figure 4).  The distribution of brown-down date in each year is slightly different from each other as well.  This trend of brown down date show a climatic factor or local environment influence other than latitude affecting brown-down phenology.  The data are not well classified in 2006 because of a high degree of aggregation of 284 Julian date. 

Figure 4 Brown-down date of aspen from the west Canada to Minnesota (US) based on EVI data from 2001 to 2006.  In 2006, the brown down date is more frequently inferred as 284, which may be due to unclassified pixels of EVI images. 

The map series of brown-down in aspen forest are provided as below. 

Daylight length as a trigger

Daylight length is a important trigger of brown-down or bud set.  Figure 5 and Figure 6 show the relationship between the phenological event and daylight length as a trigger. 

Figure 5 Scatter plot of brown-down date versus daylight length.  

Figure 6 Regression for date of bud set in 2000 versus daylight length of provenance origins in the Athabasca trials.  Each point is the square of bud set score in each seedlot recorded on September 14th, 2000, which is one day after the daylength records.  The relation is significant at the 0.05 level.
 

Figure 4 is a lattice plot of brown-down date versus environment variables (Longitude, NFFD, elevation and daylight length).  The orange solid line is the panel level regression line.  The legend shows different provenances which are the same as in the plots below.  The Minnesota seed sources  (in green) and the northern BC, Alberta provenances (pink) are different from other ones in terms of the adapted autumn temperature and the end of growing season.   The relationship between the mean annual temperature and the bud set score is weaker than other three climatic variables.  

Climatic factors

Brown-down date versus climatic variables

The figure following is a lattice plot of brown-down date versus climatic variables (FFP days, MAT, DD>5°C and eFFP).  The color legend shows different provenances which are the same as above. The Minnesota seed source (in green) is different than other provenances.  The relationships between brown-down and climate variables are significant compared with bud set versus environmental variables. 

Figure 3 is a lattice plot of brown-down date versus environmental variables (Longitude, NFFD, daylight length, and latitude).  The green points BS is the boreal shield zone in Minnesota. Black ones are from  BP zone (Boreal Plain in Alberta and Saskatchewan ). The pink ones are from TP zone, the north of Alberta and the northwest of BC.  The red color indicates FH zone, the foothill in Alberta.  

Figure 7 Scatter plots of brown-down date versus climatic variables. 

Bud set date versus climatic variables

Figure 8 is a lattice plot of bud set date versus climatic variables (TAV_at, TAV_August, eFFP and MAT).  The color legend are the same as above. The Minnesota seed source (in green) is different than other provenances in the scatter plots.  The relationship between bud set and elevation is less significant than the relationships between brown-down date versus climatic variables.  

Figure 8 Scatter plots of bud set date versus climatic variables. 

Table 1 Pearson correlations (r) between geographic, climatic and brown-down dates (annual date from 2001 to 2006 and the average).  The absolute values of red shaded coefficients are more than 0.80.