Geology Reference
In-Depth Information
Model unable
to exploit data
Identifiability
problems
Fig. 20.1 Schematic diagram of the relationship between model complexity, data availability and predictive
performance (reproduced with permission from Grayson & Bloschl, 2000). The added arrows are: the full line
which is the ideal path of model development where added complexity is supported by new data; and the
dashed line which is the more common scenario where new data acts to better constrain existing models.
dashed arrow: more often than not we need more
and better data rather than more complex models
to improve erosion predictions.
And so we come back to the subject of the
chapter and ask: 'In what direction will the future
of soil erosion modelling go?' The 1950s to the
1970s was the period of empirically-based erosion
equation development, culminating in the sec-
ond Universal Soil Loss Equation release in 1978
(Wischmeier & Smith, 1978). The 1970s to the
1990s was a period of development of process-
based erosion models facilitated by computer
simulation modelling. As discussed above, much
of the progress during the 1990s and 2000s has
focused on understanding and representing uncer-
tainty associated with model applications. In
most recent years we have seen, and (we predict)
into the future we will continue to see, advances
in spatial modelling and up-scaling, interfaces
that utilize GIS, the increased use of remotely
sensed data, and web-based delivery systems tied
to large databases for soils, topography, land use
and weather. This does not mean that we will not
also see advances in other aspects of the model-
ling. Empirical modelling has continued, for
example, as evidenced by the publication of the
Revised Universal Soil Loss Equation in 1997
(Renard et al ., 1997), and there is much work yet
to be done on solving the problems associated
with model and data uncertainty, as discussed
above. Uncertainty will continue to play a large
role in our thinking on erosion, and ideally will
lead to new ways of both modelling soil erosion
and thinking about how we manage land for pur-
poses of conservation. Nonetheless, practical
goals associated with management decisions will
require a set of model requirements that stress
data reliability and availability, ease of use, capa-
bility for routing water and sediment through
watersheds, and ability to delineate primary
trends in erosion rates as a function of manage-
ment practices and changing climate.
Search WWH ::




Custom Search