Civil Engineering Reference
In-Depth Information
Where calibration is limited or not possible due to a lack of measured data,
a sensitivity analysis is often used to estimate the uncertainty in the model.
This involves varying model parameters around the estimated parameter
value within a range that is considered to be feasible to assess the extent to
which model outputs are affected by the change. Results give one indication
of the uncertainty in the model results.
Future scenarios for the water resource
Because water resource plans are forward looking, a prediction of future rainfall,
inflow or recharge is a requirement. When combined with water resource
models this enables future available water to be estimated. The most common
approach used up until recently has been to assume future rainfall and river
flows would be similar to available historic records of rainfall and river flows
(typically from the previous twenty to one hundred years). For static models
this means using long-term statistical figures derived from historic infor-
mation, such as annual average, median, 90 percentile dry, or driest on record.
More recently there has been greater consideration of future climate
scenarios outside of the range of recorded historic data, to reflect evidence of
climate change or longer term climatic cycles beyond the period of recorded
historic rainfall and river flow records. Scenarios are different from forecasts:
scenarios explore a range of possible outcomes resulting from uncertainty;
whereas forecasts identify the most likely pathway and estimate uncer-
tainties. Therefore, 'scenario planning is not forecasting of the most probable
future but it creates a set of the plausible futures' (Amer
et al
. 2013: 25).
Climate change offers an additional level of uncertainty, with the IPCC
concluding that 'observational records and climate projections provide
abundant evidence that freshwater resources are vulnerable and have the
potential to be strongly impacted by climate change, with wide ranging
consequences' (IPCC 2008). Lower rainfall in places will likely mean more
sunshine, higher temperatures, and therefore higher evapo-transpiration
rates. At a time of increased irrigation demand from storages and rivers, it
also leads to less surface runoff to rivers at the same time as the environment
is also water-deprived. Faurès
et al
. (2010) report that small variations of
rainfall usually translate into much larger variations in river runoff. They
referred to an example of analysis of historical precipitation and runoff data in
Cyprus that showed that an average reduction of 13% of annual precipitation
translated into a 34% reduction in runoff. Likewise in the Murray-Darling
Basin in Australia, a 13% decline in rainfall observed in the southern MDB
from 1997-2006 (which included drought years) resulted in a 44% decline
in streamflow (Horne 2013: 142). In terms of modelling future flows in the
Ebro basin in Spain, Garcia-Vera (2013: 226) reports that:
For the 2010-2040 timeframe, a 12% average decline for the Ebro basin
was estimated, or what is equivalent to a 17% decrease in surface flows,
