467
them ideal for the exploration of a large number of scenarios. This is particularly
valuable for two-way interactions and co-development with decision-makers.
In response to the need for regional or local integrated assessment, recent
research has focused on including variable spatial resolution, or a “telescoping”
ability. This is a comparable effort to the capability for multiple scales in Earth
System models. The capability has already been used to disaggregate the US energy
and economy systems into 50 states, and a province-level version of an IAM has
recently been created for China.
However, this variable-spatial-resolution capability has not yet been applied to
explore sub-state level FEW system dynamics. Place-based research can be used as
a means to take on several critical challenges that present themselves in applying
IAMs to address sub-state resolution of FEW systems (e.g., major urban centers,
sub-basin representation and major water demand sectors, grid-level agricultural
representations).
One important modeling challenge in this regard is the representation of hydrology. IAMs typically use a gridded global hydrologic model with a fixed spatial resolution and allocate water demands among the competing water users in river basins
globally. This scale (in the order of 0.1–0.5°) is large enough to limit the effects of
between-basin water flows, except for major water transfer projects.
At finer spatial scales, it will be critical to represent the linkages among subbasins, i.e., upstream-to-downstream dynamics, which implies major advances in a
reduced form hydrology modeling approach that has thus far not needed to include
these water transfers. This telescoping over a much smaller scale than traditionally
applied in IAMs will also dictate investigating the scalability of the current representation of the hydrologic processes (e.g., runoff generation, river) to finer spatial
scales. Future efforts in FEW system modeling using IAMs can achieve this by
incorporating interbasin water flow capabilities within the hydrology model or
switch to a finer resolution model (e.g., SWAT).
Second, the successful development and use of modeling tools for decisionmaking require a two-way process in which decision-makers are co-developers with
the modeling teams. Questions about when and where higher resolution is important or beneficial are dependent critically on the specific decisions and decision
processes to be supported by modeling. This co-development, thus, requires the
creation and strengthening of social networks that include decision-makers and
analysts.
This challenge involves the representation of human decision-making, which
applies across the FEW system nexus. In the case of IAMs, economic decisions are
made using a discrete choice formulation (comparing two or more discrete alternatives, e.g., McFadden 1980). At sufficiently large spatial scales, there exists sufficient information to calibrate these discrete choice functions. However, at finer
spatial scales, information about historical choices may be insufficient to calibrate
the tools, particularly under the assumption of major new technology options or
price changes.
A working hypothesis is that these discrete choice models can be used for finerscale representations, but that doing so will require innovative methods that utilize
17 Applying Science to Practice
them ideal for the exploration of a large number of scenarios. This is particularly
valuable for two-way interactions and co-development with decision-makers.
In response to the need for regional or local integrated assessment, recent
research has focused on including variable spatial resolution, or a “telescoping”
ability. This is a comparable effort to the capability for multiple scales in Earth
System models. The capability has already been used to disaggregate the US energy
and economy systems into 50 states, and a province-level version of an IAM has
recently been created for China.
However, this variable-spatial-resolution capability has not yet been applied to
explore sub-state level FEW system dynamics. Place-based research can be used as
a means to take on several critical challenges that present themselves in applying
IAMs to address sub-state resolution of FEW systems (e.g., major urban centers,
sub-basin representation and major water demand sectors, grid-level agricultural
representations).
One important modeling challenge in this regard is the representation of hydrology. IAMs typically use a gridded global hydrologic model with a fixed spatial resolution and allocate water demands among the competing water users in river basins
globally. This scale (in the order of 0.1–0.5°) is large enough to limit the effects of
between-basin water flows, except for major water transfer projects.
At finer spatial scales, it will be critical to represent the linkages among subbasins, i.e., upstream-to-downstream dynamics, which implies major advances in a
reduced form hydrology modeling approach that has thus far not needed to include
these water transfers. This telescoping over a much smaller scale than traditionally
applied in IAMs will also dictate investigating the scalability of the current representation of the hydrologic processes (e.g., runoff generation, river) to finer spatial
scales. Future efforts in FEW system modeling using IAMs can achieve this by
incorporating interbasin water flow capabilities within the hydrology model or
switch to a finer resolution model (e.g., SWAT).
Second, the successful development and use of modeling tools for decisionmaking require a two-way process in which decision-makers are co-developers with
the modeling teams. Questions about when and where higher resolution is important or beneficial are dependent critically on the specific decisions and decision
processes to be supported by modeling. This co-development, thus, requires the
creation and strengthening of social networks that include decision-makers and
analysts.
This challenge involves the representation of human decision-making, which
applies across the FEW system nexus. In the case of IAMs, economic decisions are
made using a discrete choice formulation (comparing two or more discrete alternatives, e.g., McFadden 1980). At sufficiently large spatial scales, there exists sufficient information to calibrate these discrete choice functions. However, at finer
spatial scales, information about historical choices may be insufficient to calibrate
the tools, particularly under the assumption of major new technology options or
price changes.
A working hypothesis is that these discrete choice models can be used for finerscale representations, but that doing so will require innovative methods that utilize
17 Applying Science to Practice