Agriculture Reference
In-Depth Information
effect of water transformation; (iii) a C/N/P
dynamics and structure model, CAST, that
links the transformations of organic matter
with a dynamic model of soil aggregation/
disaggregation, a simplified terrestrial ecol-
ogy model that is comprised of mycorrhizal
fungi, microorganisms (BIO pool), consumers
and predators (FAUNA pool) and a plant/
root dynamics model; (iv) a plant dynam-
ics model, PROSUM, that is based on bio-
mass production ecological principles
and predicts the dynamics of key variables
(e.g. above- and belowground production
of litter C and N; nutrient and water up-
take) in response to key drivers (tempera-
ture; availability of light, water, CO
2
and the
nutrient elements N, P, Ca, Mg and K; grazing
and management events) for the wide range
of vegetation types (outlined in Banwart
et al
., 2012).
Integrating a
1-
D ecosystem function
model into a fully distributed catchment
model would provide a rigorous biophysi-
cally based model that computes rates of
change and updates numerous state vari-
ables based on biophysical principles
within the soil profile, while transporting
the water and solutes to streams draining
the catchment.
The conceptual strategy goes beyond a
distributed C-N-biogeochemical model
that integrates multiple data sources to
include a database that links the model
to the 'essential terrestrial variables', or
ETVs, that allow dynamic simulation across
spatial and temporal scales relevant to
decision making for soil management and
land use. As part of this model synthesis
is the question of data and data support.
A multi-scale model as proposed here
must also have efficient access to data for
topography, soils, geology and climate,
land cover/land use, as well as reactive
transport properties and parameters and
open-source GIS tools to set up the model
domain and parameters. Dynamic model-
ling of geospatial landscape processes re-
quires a comprehensive database that can
be linked to each land use, soil, geology
and climate. This model-data architecture
is a major challenge and, to our knowledge,
existing databases lack the harmonized data
products necessary for modelling that
reflect current and past land-use and
management practices, are geospatially and
geotemporally consistent and are fully
coupled to the model within an efficient
cyberinfrastructure. Such a prototype is
being developed at PennState (see
http://
www.hydroterre.psu.edu
) and at the EC
Joint Research Centre (see
http://fate.jrc
.
ec.europa.eu/interactive- maps-and-data).
Figure 8.4
presents a prototype service
showing the soils, land parcel and wetland
inventory in a central database used to
automate the development of the PIHM
model for a site in Pennsylvania, USA.
Ideally, this prototype cyberinfrastructure
would enable users to have efficient access
to all necessary data products and to be
able to carry out simulations for any catch-
ment in a region from a consistent data
source.
Clearly, coupling a catchment hydro-
dynamic model with an integrated carbon
and nitrogen and reactive transport code
that resolves processes at spatial scales from
plot to catchments must overcome several
obstacles:
•
Data support across scales
: the process
of connecting an independent C-N re-
active transport model with a spatially
distributed hydrodynamic model will
require a new approach to model data
and parameterization. We propose a
new strategy that integrates geospatial/
temporal data, political or property
boundaries, hydrographic, hydrologic,
soil, climate, reactive transport proper-
ties and topographical and other com-
munity-derived geospatial data (
http://
•
Multi-state parameter estimation
: a wide
range of parameter estimation tools are
presently available that can be adopted
for automated parameter estimation.
•
Interoperability
: the lack of interoper-
ability between water data, models
and model simulations, and the lack of
fast access between essential water-
shed data and computing resources,
Search WWH ::
Custom Search