Geoscience Reference
In-Depth Information
is difficult to manipulate the inputs for a whole catchment (but see the Gardsj on catchment experiment
Rodhe
et al.
, 1996). Thus catchment scale analyses are generally limited to environmental tracers and
particularly the isotopes of oxygen and hydrogen/deuterium/tritium (
18
O/
16
O,
2
H/
1
H and
3
H) that form
part of the water molecule and should consequently be good tracers of water flow pathways. Tritium is
radioactive and mostly derives from atmospheric nuclear bomb tests in the 1950s and 1960s. Such tests
were banned in 1963 (although France continued testing until 1974 and China until 1980), so that the
levels of tritium in the atmosphere have since declined, accelerated by radioactive decay of the remaining
tritium (which has a half life of 12.32 years). Until relatively recently, isotope analyses were expensive
and time-consuming so that many other chemical characteristics have been used to infer water sources
and pathways from the early study of Pinder and Jones (1969) onwards. A new generation of laser spec-
trometers is now bringing down the cost of isotope analyes (e.g. Berman
et al.
, 2009) and we can expect
to see many more intensive studies of isotope behaviours in hydrological systems in future. It has been
argued in a number of papers that tracer information, where it is available, can provide useful additional
constraints for use in model calibration to test whether rainfall-runoff models are getting the right results
for the right reasons (e.g. Seibert and McDonnell, 2002; Kirchner, 2006; McGuire
et al.
, 2007; Vache
and McDonnell, 2006), although the need to reproduce tracer concentrations as well as discharge hydro-
graphs generally means that more model parameters need to be calibrated. This is discussed further in
Chapter 11.
3.11 Linking Model Components and Data Series
There are increasing demands to link rainfall-runoff models to other model components that might be
concerned with sediment transport, water quality, hydroecology, urban drainage control systems, etc.
In the past, this has tended to be done within specific modelling packages, though there were some
early attempts at a more general solution, such as the ANNIE database system used by the USGS
(Leavesley and Stannard (1995) include a brief description in their work) to link hydrological mod-
els to spatial and temporal data files, and the UK Institute of Hydrology Water Information System
(WIS). In the last decade, there have been significant developments in working towards standards
for such exchanges (and not only in hydrology - this is a problem common to many environmental
modelling domains).
One such major initiative is the Open Geospatial Consortium (OGC) which has worked since 1994
to define standards for geospatial data and services. The US-based Consortium of Universities for the
Advancement of Hydrologic Science Inc. (CUAHSI, www.cuahsi.org) has been developing open source
standards for exchange of data in its WaterML and CUAHSI Hydrologic Information System (CUAHSI-
HIS) that are intended to be consistent with the OGC standards. The European Community has also
funded a number of projects of this type. The OpenMI Association (see www.openmi.org/reloaded) is
developing open source software designed to provide common interfaces between models and data on a
time step by time step basis during simulations (see Gregerson
et al.
, 2007). If model components are
programmed to be compatible with the interface definitions then different components should be able
to automatically exchange information, regardless of the spatial and temporal time steps. The standards
support distributed model components with different grid sizes. The Interoperability and Mapping project
(INTAMAP, www.intamap.org) is developing a web-based framework for the real-time mapping of
critical environmental variables using geostatistical methods. It can be used to exchange data, carry
out statistical analysis and visualise results. OGC-compatible standards for communicating uncertainty
in variables and model results are also being developed using UncertML (www.uncertml.org). This is
a rapidly developing area that is also taking advantage of developments in
cloud computing
, such as
GoogleMaps and the Google Earth Engine, that allow the presentation of model results in a highly visual
form (Figure 3.7).
