| dc.description.abstract | Ecologically based management of white spruce (Picea glauca [Moench] Voss.)
requires an understanding of its patterns of adaptive variation. Six common garden trials
and a greenhouse trial established in 2002 and 2003 across Ontario were used to assess
levels of genetic variation in 127 seed sources from Ontario and western Quebec and
relate this variation to local climate. Using this information focal point seed zones were
developed. The focal point seed zone methodology determines spatially explicit areas of
ecological compatibility for any selected point. This approach will assist in properly
matching seed sources and planting sites based on current and predicted future climate
conditions.
Growth and phonological variables, including height, root collar diameter,
survival, budflush timing, and budset timing were measured. Intraclass correlation
coefficients were calculated for all traits to determine levels of genetic variation. Levels
of between-provenance genetic variation ranged from 0 percent for several of the
budflush variables, up to 22 percent of the total amount of variation expressed for 2003
survival at the Englehart field trial. Overall, growth variables showed higher levels of
between-provenance variation than phonological variables. Simple linear regressions
were used to relate these differences to local climate conditions. Variation was explained
by a wide range of temperature and precipitation related variables. Late budset stages,
which had r2 values ranging from 0.55 to 0.46, were explained by temperature and
precipitation variables related to the growing season. Generally, the primary patterns of
adaptive variation followed a southeast to northwest trend across Ontario. A secondary
east-west trend was evident in northwestern Ontario. Northern sources flushed earlier
and set bud earlier, while southern sources demonstrated superior growth. Results
support previous white spruce genecology studies showing superior growth of sources
from the Ottawa Valley region of Ontario and Quebec.
Two statistical approaches were used to develop focal point seed zones. The first
used principal components analysis (PGA) to summarize patterns of variation based on
selected variables. Provenance factor scores were then regressed against climate
variables and the resulting equations used to model the PC axes. The second approach
used canonical correlation analysis (cancorr) to simultaneously find the relationship
within and between biological and climate data sets. Standardized climate coefficients from each significant canonical variate were used to model patterns of adaptation. For
both methods parallel seed zones were constructed using GIS tools to intersect grids
standardized to sample points selected from across the study area. Results showed
overall similar trends for the two methods, however, cancorr based zones showed
stronger longitudinal trends for northern points and became more fragmented for
southern points. Cancorr zones were also more affected by lake shore effect from Lake
Superior and Georgian Bay than regression based zones. | |
