|
1 |
Architecture
A framework Consisting of Simulation Models, Inputs, Analytical outputs, and Web application .
|
M |
Modeling Application System Interactive framework
(MASIF) whose components are shown in this figure is characterized by: (1) a
scalable data management module for rapid and ready access to input and output
data; (2) a visualization module for the exploration, description, and analysis
of spatial and temporal patterns; (3) a statistical analysis module to conduct
and compare model scenarios; (4) an output animation module to produce spatio-temporal time series of model output; (5) Web based
interface to interact with the model (Fig. 1).
Fig. 1 Integration of Models and input output analysis.
Framework. Masif includes six software products: Visual Basic 6.0,Oracle 8.05, MS Access 2000, ArcView
3.2, MineSet 3.0, and S-Plus 2000 2.0. These products
represent a class of existing upgradeable applications that are inherently useful
for the analysis of large data sets, are widely used worldwide, and include libraries
that facilitate interconnections. The major steps in the implementation of MASIF included: (a)
the design of database tables for data storage and management; (b) the
development of the interface to the geographic information systems (GIS)
module; (c) the establishment of connections between the Visual Basic
interface and the analytical software; and (d) the development of a user interface
to integrate a specific model to the various software options (Fig. 2).
Fig. 2 MASIF software products and their connectivity.
Databases.
Database tables contain the data infrastructure for MASIF: Input
weather variables, soil variables
tables and output variables tables, and a parameters table. The process of
using a parameters table allows users to store parameters during different
model runs rather than storing millions of records of model output data
associated with parameter modifications. Users can recall the parameters used
for a specific simulation run and re-run the model.
Mapping.
Geographic information system (GIS
) module. To
perform in-line mapping of simulation results, we integrated a geographic
information system into MASIF by developing a set of scripts to
access model output, conduct spatial
interpolation of selected output variables at any place in
time, and present the resulting map(s) on the computer display. The
scripts were developed using ArcView Avenue. In addition, modelers have access to the
ArcView interface which
allows them to optimize other GIS analytical tools
provided by the software.
Connectivity. The connection between ArcView
and Visual Basic is established using DDE, an MS
Windows supported client/server mechanism that
enables two applications to communicate. When the user selects the GIS option,
the interface (Visual Basic) initiates communication with ArcView
and sets it up as the server. Visual Basic thus becomes the client and
requests ArcView to implement
specific tasks by issuing Avenue commands through the Dynamic Data Exchange (DDE) channel. Through Data Dynamic Exchange
(DDE), Visual Basic launches ArcView, closes the
default ArcView project, opens the project containing
the scripts and calls the master script with the user specified model parameters as input. Thereafter, the master script calls
the necessary scripts within the project to produce the analysis specified by
the user.
The connection between MineSet and Visual Basic was developed using ActiveX technology. MineSet provides an ActiveX control, named VizComposite, that
is able to display any type of MineSet visualization
within a Visual Basic application. When the user selects the Multidimensional
Visualization option, a MineSet schema file is
created based on the user input and passed to the VizComposite
control. The control then interprets the schema file appropriately and displays
the visualization within Visual Basic.
The connection between S-PLUS and
Visual Basic was made using an Object Linking and Embedding (OLE) Automation,
which enables one application (client) to access the resources and
functionalities of another application (server). In MASIF, S-PLUS is the
application server that provides resources in the form of Type Libraries, a
collection of objects, functions, properties and methods. The interface (Visual
Basic) is the application client that calls the appropriate
Type-Library
entities to execute the analysis requested by the user. Upon the execution of a given S-PLUS command, the resulting
graph-sheet or report is returned to Visual Basic as an OLE object and is thus
embedded into MASIF.
Simulations Process
A computing
process that runs the desired Model and provides the input weather and soil
data to generate outputs for a selected county, all North central region, Resac region with the appropriate parameters chosen. If the
model requires weather data a database weather table provides 1055 daily
weather parameters (Fig. 3).
Fig. 3 Simulation
Process
Analysis Process
A
computational and mapping process that converts the generated model outputs
into GIS mapping images through Raster generation step and subsequently shows
such images to the users for analysis and interpretation through Raster
Display, Model Comparison, Actual Comparison, and Animation (Fig. 4,5,6,7,8,9,10).
Fig. 5. Analysis Process Steps
Fig. 6 Analysis - Raster generation step.
Fig. 7
Analysis Raster Display.
Fig. 8 Analysis Model comparison.
Fig. 9 Analysis Actual comparison.
Fig. 10 Analysis Animation.
Web Application
A Web
application has been developed using the final Models outputs. The yearly and
daily analysis of Corn and Hybrid Maize have been
completed and other models are under development. Corn and Hybrid Maize outputs
are displayed through line charts, bar charts, and the regional maps that are
generated separately by models. At the bottom of each output page weather data
menus were implemented to provide a history of the weather variables data for
the selected county. In addition, a regression analysis image output for the
displayed region was presented. The MASIF web site also provides a 1055 county
regional actual graph data analysis to show the 30 years history of corn
production fluctuations. The link to the web is: http://masif.cevl.msu.edu/models.asp
|
Chapter 2 |
Corn Model Sinclair - Muchow
A daily Simulation model of growth and production of corn.
|
M |
uchow model simulates the corn growth and production on a
daily weather input. The input variables consist of five variables for 1055
locations amounting to 11.9 million records, was
structured in a relational database management system using Oracle. The
Analysis step generates Mapping images from the simulated output database.
Simulation
To Run Muchow simply click on the MASIF icon, and Model and Muchow buttons. From the Model Muchow
Linear Calculation screen select Year, Region, Parameters, and Available
Water. From the year menu select one, several, or all the years. From the
region menu you have the option to select one, several, All
counties, or Resac counties. Click on the Run Model
button and wait until the simulation is completed (Fig. 1).
Fig. 1
Model Muchow Linear Calculation
When simulation
is completed then by following the steps in the Analysis section will generate
the map. In Raster Generation from the MODEL and SIM.
NO. menu select Muchow and the
simulation number and then Parameters, years, and days. Click on the Execute
button to generate the map, Raster Display will show the map on the screen,
Model Comparison compares the output maps from the two models with Chi-Square statistics
and subtractions. Actual comparison compares the model with Chi-Square
statistics and subtractions. Animation shows the animation production image progress
for a particular year (Chapter 8).
Database Tables
·
Input NEWWEATHER
·
Input MODELTRANSACTION
·
Output MUCOUTPUT
Flowchart
Muchow Corn simulation model flowchart and input output processes. It starts
by reading from model transaction and weather tables parameters and looping
through years and regions and calling the Corn executable program and updating
the output tables.
Fig. 2 Corn Muchow
simulation flowchart.