Geography Reference
In-Depth Information
3 Regional Scale Visualisation
Visualisation of the duration and extent of projected changes in the Victorian
climate was achieved through two complementary approaches:
• The creation of an animation providing an overall statewide perspective of the
extent of the projected changes in temperature and rainfall and rate at which
these changes are likely to occur.
• The creation of tables and graphs associating each centroid point of cells from a
0.05 degree Victoria raster map with its projected monthly and annual average
temperatures and rainfall across multiple years.
3.1 Temporal Animation Development
Animations depicting the duration and extent of projected changes in annual
average temperature and rainfall, across the state of Victoria, between the year
2000 and 2050 were developed through the following workflow:
Annual average values for temperature and rainfall were calculated in ESRI
ArcGIS 9.3 for the year 2000 and 2050 using monthly climate records and climate
change projections respectively. In order to create a visually smooth transition
between the dataset corresponding to the year 2000 and the projected data for
2050, synthetic data was calculated through a backward propagating method from
2050 using a yearly time step. This was done to avoid the display of unrealistic
sharp transitions between the data, and assumed a linear temperature increase from
one year to the next. The applied method could easily be altered to take into
account intermediate years as they become available, or as derived from a different
interpolation formula. In creating such a data visualisation product a certain
amount of artistic license is available and corresponding care is required.
As reported by Monmonier (
1996
), all cartographic products including data vi-
sualisations may include a number of cartographic 'white lies'. What is most
important is that an ethical stance is taken by the visualisation producer to depict
and convey the correct context and meaning to the output, as reported by Sheppard
and Cizek (
2009
). In this case, the scale of the changes was felt to be more
important than the specific trajectory, especially as data for interpolation was not
easily available.
The climate raster data for the years 2000, 2050 and intermediate years was
then exported to Keyhole Markup Language (KML) as a series of spatially
projected image overlays, using geoprocessing scripts previously developed by
Aurambout and Pettit (
2008
). Next, we made use of the KML 'time-span' func-
tionality to automatically specify the dates and duration for which each overlay
should be visible in Google Earth. The use of this functionality enabled the display
