Spatial Modelling and Scaling Issues - Environmental Modelling: Finding Simplicity in Complexity - page 79

parameters and coarse-resolution predictor variables.
Second, weather-pattern approaches of downscaling typ-
ically involve statistically relating observed station or
area-average meteorological data to a given weather clas-
sification scheme, which may be either objectively or
subjectively derived. Third, stochastic weather generators
are designed to simulate daily time-series of climate data
for each successive day governed by the characteristics of
previous days based on first or multiple-order Markov
renewal processes. These models are commonly used
for climate-impact studies. Fourth, limited-area climate
models are the high-resolution limited-area climate mod-
els embedded with GCM and use the GCM to define their
boundary conditions. These models have the ability to
simulate small-scale atmospheric features and may pro-
vide atmospheric data for impact assessments that reflect
the natural heterogeneity of the climate at regional scales.
There are also many other techniques, including climate-
sensitivity analysis of impact models, construction of
spatial/temporal analogues and the simple change-factor
technique (Wilby et al ., 2004).
The change-factor technique is the simplest and fastest
approach for processing numerous data sets and feasi-
ble for global-scale downscaling of climate simulations
(Wilby et al ., 2004). For example, Tabor and Williams
(2010) have implemented the change-factor technique to
downscale many climate simulations for the twenty-first

century from the Intergovernmental Panel on Climate
Change Fourth Assessment Report (IPCC AR4)(IPCC,
2007). Briefly, the change-factor approach first establishes
the reference climatology and interannual variability for
the site of region of interest (Wilby et al ., 2004; Tabor
and Williams, 2010). The reference climatology is a rep-
resentative long-term average or an actual meteorological
record. Secondly, it calculates the changes in the equiv-
alent climate variable for the GCM grid-box closest to
the target site. Finally, the climate change suggested by
GCM is then simply added to each month (or day) in the
reference climatology. Using the change-factor approach,
Tabor and Williams (2010) have downscaled tempera-
ture and precipitation projections for the 21 st century
from GCM model data (3 ◦ ) to 10-minute resolution
(Figure 5.9).

5.6 Scaling land-surface parameters for a
soil-erosion model: a case study

Soil erosion is recognized as amajor problemarising from
agricultural intensification, land degradation, and possi-
ble global climatic change (see also Chapters 15 and 23).
However, the extent of the problem is hard to quantify as
field measurements of erosion are rare, time-consuming,
and are only acquired over restricted temporal and spatial

Publisher's Note:
Image not available
in the electronic edition

Figure 5.9 An example indicating the spatial patterns of annual mean temperature and precipitation at 10 minutes downscaled from
the GCM data with native grid resolution of 3 ◦ . The A1B HadCM3 is a GCM scenario of Hadley Center for Climate Prediction (UK)
(Reproduced with permission from Tabor K., and Williams, J.W. (2010) Globally downscaled climate projections for assessing the
conservation impacts of climate change. Ecological Applications , 20 (2), 554wnsc).