|
|
Error and Uncertainty
Any study of climate change on sub-continental level is limited by the temporal extant of the data available.
Climatic shifts are observable as trends superimposed onto natural variation. A study of only 30 years is a totally
inadequate in making conclusions as to changes across BC. This project is based on the findings of other larger projects
and makes the assumption that global climate is occurring. Adequate Data is available for BC Columbia for a further
70 ¡V 80 years and with more time this could be used to provide a century-wide view of minimum temperatures.
Without a long-term data set we cannot be sure weather the unusually mild period we experienced recently is an
indication of warming temperatures or is a part of natural variation. Studies using historical data can not do this,
paleoclimatic data is needed.
The distribution of weather stations does also impact the validity of our results, particularly in
Northern regions where many 100 hundreds of Km^2 are represented by a handful of stations. A large amount of
variation across the province may have been overlooked in the interpolation.
Using elevation (altitude) as the key variable allowed us to make a simple interpolation of minimum temperatures
across the study area. This however was a major assumption as other impacts influence temperature, such as latitude,
atmospheric circulation patterns, local conditions, vegetation, continental, and ocean current characteristics along
coastal locations.
Using the interpolation using location at sea level did go some way to alleviate this problem. However at
high latitudes where there were few stations this caused inaccuracies. The lowest temperature ever recorded in British
Columbia was -58 C 1 in 1947. Our interpolation suggests that this temperature was
regularly reached and exceeded in Northern BC which indicates that we have overestimated the minimum temperatures
in these regions, possibly because very low temperatures were recorded at relatively low attitudes and then extrapolated
to high attitudes. Temperatures in Coastal and Southern BC were consistent with expected extremes.
Handling then data in annual years rather than seasons may have influenced results. Each map handles data
from the end of one winter and the start of the next. Handling the season together would produce more coherent results.
The cold events for event station tended to fall in the same month so that the cold event from one season could be lost
if it fell at the end of winter and the following season cold event occurred at the start of the following season.
The 1 Km resolution was an appropriate size considering the size of our study area in terms of accuracy and
data management. A repeat of this study would be improved by reducing the study area to the Southern Half of BC,
particularly the interior plateau, and using a DEM with a higher resolution.
Reference
1 North American Weather Almanac
http://www.islandnet.com/~see/weather/almanac/arc2002/alm02feb.htm
|