404 lines
11 KiB
HTML
404 lines
11 KiB
HTML
<!DOCTYPE html PUBLIC "-//w3c//dtd html 4.0 transitional//en">
|
|
<html>
|
|
<head>
|
|
<meta http-equiv="Content-Type"
|
|
content="text/html; charset=iso-8859-1">
|
|
<meta name="GENERATOR"
|
|
content="Mozilla/4.79 [en] (X11; U; Linux 2.4.18-27.7.xsmp i686) [Netscape]">
|
|
<title>gfeConfig - Algorithm Configuration</title>
|
|
</head>
|
|
<body style="background-color: rgb(255, 255, 255);">
|
|
<div class="Body">
|
|
<div class="3Heading">
|
|
<h2 style="text-align: center;">GFE/ifpIMAGE Configuration File<br>
|
|
Algorithm Configuration<br>
|
|
</h2>
|
|
</div>
|
|
<br>
|
|
<table nosave="" border="1" cols="4" width="100%">
|
|
<caption> <br>
|
|
</caption><tbody>
|
|
</tbody><tbody>
|
|
</tbody><tbody>
|
|
</tbody> <tbody>
|
|
<tr>
|
|
<td>
|
|
<center><b>GFE Configuration Item</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Applicable to GFE</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Applicable to <a href="ifpIMAGE.html"><!--disable indexer-->ifpIMAGE</a></b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Notes</b></center>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="#Significantweathertimeweightedaveragepercent">Significant
|
|
Weather Time-Weighted Sample Threshold</a></td>
|
|
<td>YES</td>
|
|
<td>NO</td>
|
|
<td><br>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="#DefaultPencilWidth">Default Pencil Width</a></td>
|
|
<td>YES</td>
|
|
<td>NO</td>
|
|
<td><br>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="#PencilInfluenceSizes">Pencil Influence Sizes</a></td>
|
|
<td>YES</td>
|
|
<td>NO</td>
|
|
<td><br>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="#InterpolationAlgorithm">Interpolation Algorithm</a></td>
|
|
<td>YES</td>
|
|
<td>NO</td>
|
|
<td><br>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td valign="top"><a href="#SmoothingAlgorithm">Smoothing Algorithm</a></td>
|
|
<td valign="top">YES</td>
|
|
<td valign="top">NO</td>
|
|
<td valign="top"><br>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="#ContourSubSample">Contour Sub Sample</a></td>
|
|
<td>YES</td>
|
|
<td>NO</td>
|
|
<td><br>
|
|
</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
<blockquote>
|
|
</blockquote>
|
|
<hr style="width: 100%; height: 2px;">
|
|
<h4><br>
|
|
</h4>
|
|
<h4><a name="Significantweathertimeweightedaveragepercent"></a><font
|
|
size="+1">Significant
|
|
weather time weighted average percent</font></h4>
|
|
Smart tools can request a time-weighted average of grids. For
|
|
weather,
|
|
this requires a threshold. The weather key returned from the
|
|
time-weighted
|
|
average of grids, includes all weather sub keys that exist for more
|
|
than
|
|
the Significant WeatherTimeWeightAverage_percent of the grids. If
|
|
none of the sub keys exceed the threshold, then the most dominent
|
|
subkeys)
|
|
will be used as the time-weighted average.
|
|
<p>For example, let's say there are three weather grids, the first is
|
|
an
|
|
hour long, the second is two hours long, and the third is three hours
|
|
long.
|
|
There is Rain in the first grid, Snow in the second grid, and Thunder
|
|
and
|
|
Rain in the third grid. The total duration of the three grids is
|
|
6 hours. The percentages calculated via this algorithm is shown in the
|
|
following table:</p>
|
|
<br>
|
|
<table nosave="" border="1" width="100%">
|
|
<caption> <br>
|
|
</caption><tbody>
|
|
</tbody><tbody>
|
|
</tbody><tbody>
|
|
</tbody> <tbody>
|
|
<tr nosave="">
|
|
<td>
|
|
<center><b>Weather Type</b></center>
|
|
</td>
|
|
<td nosave="">
|
|
<center><b>First Grid</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Second Grid</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Third Grid</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Total Duration</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Percentage</b></center>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td>Rain</td>
|
|
<td>1 hour</td>
|
|
<td>0 hours</td>
|
|
<td>3 hours</td>
|
|
<td>4 hours</td>
|
|
<td>66.6%</td>
|
|
</tr>
|
|
<tr>
|
|
<td>Snow</td>
|
|
<td>0 hours</td>
|
|
<td>2 hours</td>
|
|
<td>0 hours</td>
|
|
<td>2 hours</td>
|
|
<td>33.3%</td>
|
|
</tr>
|
|
<tr>
|
|
<td>Thunder</td>
|
|
<td>0 hours</td>
|
|
<td>0 hours</td>
|
|
<td>3 hours</td>
|
|
<td>3 hours</td>
|
|
<td>50.0%</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
<blockquote> <br>
|
|
</blockquote>
|
|
Depending upon the threshold, the time-weighted results can be
|
|
different,
|
|
as shown in the following table:<br>
|
|
<br>
|
|
<table nosave="" border="1" width="100%">
|
|
<caption> <br>
|
|
</caption><tbody>
|
|
</tbody><tbody>
|
|
</tbody><tbody>
|
|
</tbody> <tbody>
|
|
<tr nosave="">
|
|
<td nosave="">
|
|
<center><b>Threshold</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Result</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Notes</b></center>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td>20%</td>
|
|
<td>Rain, Snow, Thunder</td>
|
|
<td>all of the weather types are above the threshold</td>
|
|
</tr>
|
|
<tr>
|
|
<td>40%</td>
|
|
<td>Rain, Thunder</td>
|
|
<td>Snow is only 33.3%, so it is not part of the time-weighted
|
|
average</td>
|
|
</tr>
|
|
<tr>
|
|
<td>60%</td>
|
|
<td>Rain</td>
|
|
<td>Only rain is above the threshold.</td>
|
|
</tr>
|
|
<tr>
|
|
<td>80%</td>
|
|
<td>Rain</td>
|
|
<td>Since none of the entries are above the threshold, the result
|
|
is the
|
|
subkey that is most predominent.</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
<br>
|
|
<b><tt>SignificantWeatherTimeWeightAverage_percent = 40</tt></b> <br>
|
|
|
|
<h4><a name="DefaultPencilWidth"></a><font size="+1">Default Pencil
|
|
Width</font></h4>
|
|
The default pencil width is specified per weather element in grid
|
|
cells.
|
|
The pencil width can be changed interactively through the GFE
|
|
GUI.
|
|
If not specified, the default is set to 5 grid cells. For those weather
|
|
elements that are not-surface based, the format is
|
|
parmName_level_pencilWidth.
|
|
<p><b><tt>T_pencilWidth = 5</tt></b> <br>
|
|
</p>
|
|
<h4><a name="PencilInfluenceSizes"></a><font size="+1">Pencil
|
|
Influence Sizes</font></h4>
|
|
The number and value of the Pencil Tool's Influence Sizes can be
|
|
defined
|
|
in the GFE config file as well. Just define a list of sizes in
|
|
grid
|
|
cells much like the Pencil Tool Width.
|
|
<p><b><tt>PencilToolInfluence_list = [2, 4, 8, 12, 16]</tt></b> </p>
|
|
<p>You are allowed to enter as many sizes as you want, but be
|
|
warned.
|
|
The resulting menu may be too long to be completely displayed on the
|
|
screen. </p>
|
|
<h4><a name="InterpolationAlgorithm"></a><font size="+1">Interpolation
|
|
Algorithm</font></h4>
|
|
The interpolation algorithm may be specified for each weather
|
|
element.
|
|
The format of the entry is parmName_interpolateAlgorithm. If not
|
|
specified, then the format defaults to that shown in the following
|
|
table:<br>
|
|
<br>
|
|
<table nosave="" border="1" width="100%">
|
|
<caption> <br>
|
|
</caption><tbody>
|
|
</tbody><tbody>
|
|
</tbody><tbody>
|
|
</tbody> <tbody>
|
|
<tr nosave="">
|
|
<td nosave="">
|
|
<center><b>Weather Element</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Default Algorithm</b></center>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td>QPF</td>
|
|
<td>CUBIC_ADVECT</td>
|
|
</tr>
|
|
<tr>
|
|
<td>Sky</td>
|
|
<td>CUBIC_ADVECT</td>
|
|
</tr>
|
|
<tr>
|
|
<td>PoP</td>
|
|
<td>CUBIC_ADVECT</td>
|
|
</tr>
|
|
<tr>
|
|
<td>all others</td>
|
|
<td>CUBIC_NOADVECT</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
<br>
|
|
The available algorithms are shown in the following table:<br>
|
|
<br>
|
|
<table nosave="" border="1" width="100%">
|
|
<caption> <br>
|
|
</caption><tbody>
|
|
</tbody><tbody>
|
|
</tbody><tbody>
|
|
</tbody> <tbody>
|
|
<tr>
|
|
<td>
|
|
<center><b>Algorithm</b></center>
|
|
</td>
|
|
<td>
|
|
<center><b>Description</b></center>
|
|
</td>
|
|
</tr>
|
|
<tr>
|
|
<td>CUBIC_ADVECT</td>
|
|
<td>Cubic spline is applied to the data points temporally.
|
|
Software
|
|
looks for areas that can be advected. The advection component of
|
|
the algorithm works best for weather elements that contain areas of
|
|
zero
|
|
value.</td>
|
|
</tr>
|
|
<tr>
|
|
<td>CUBIC_NOADVECT</td>
|
|
<td>Cubic spline is applied to the data points temporally.
|
|
Software
|
|
gradually adjusts points from starting value to ending value without
|
|
advection
|
|
considerations.</td>
|
|
</tr>
|
|
<tr>
|
|
<td>LINEAR_ADVECT</td>
|
|
<td>Linear calculations are applied to adjacent base grids
|
|
temporally.
|
|
Software looks for areas that can be advected. The advection
|
|
component
|
|
of the algorithm works best for weather elements that contain areas of
|
|
zero value.</td>
|
|
</tr>
|
|
<tr>
|
|
<td>LINEAR_NOADVECT</td>
|
|
<td>Linear calculations are applied to adjacent base grids
|
|
temporally.
|
|
Software gradually adjusts points from starting value to ending value
|
|
without
|
|
advection considerations.</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
<br>
|
|
An example of setting Temperature to use the linear and advective
|
|
capabilities
|
|
of the interpolation, you would do the following:
|
|
<p><b><tt>T_interpolateAlgorithm = "LINEAR_ADVECT"</tt></b> </p>
|
|
<p>For those weather elements that are not surface based, the format
|
|
is
|
|
parmName_level_interpolateAlgorithm. <br>
|
|
</p>
|
|
<h4><a name="SmoothingAlgorithm"></a><font size="+1">Smoothing
|
|
Algorithm</font></h4>
|
|
The Smoothing Algorithm can use a number of possible smoothing
|
|
filters.
|
|
By default, the choices are 3x3, 5x5, 7x7, or 9x9. Generally,
|
|
more
|
|
smoothing will occur with larger smoothing filter sizes. When the GFE
|
|
is
|
|
first started, the 3x3 filter size is the default, but this can be
|
|
changed
|
|
to any one of the filter sizes in the list. To change the default
|
|
add this line to your GFEConfig file.
|
|
<p><b><tt>SmoothSize = 7</tt></b> </p>
|
|
<p>In this example the smoothing filter size will default to 7x7,
|
|
when
|
|
the GFE is started. </p>
|
|
<p>The list of possible filter sizes can be configured in the GFE as
|
|
well.
|
|
To define a new list, which appears in the main menu item
|
|
GFE->Editing
|
|
Preferences, use the following syntax in your GFEConfig file. </p>
|
|
<p><b><tt>SmoothSizeList = [3, 5, 7, 9, 11]</tt></b> </p>
|
|
<p>Note that all of the integer values are odd. Even values
|
|
will
|
|
give the same results. For example, a value of 2 will give the
|
|
same
|
|
results as 3, 4 the same as 5, and so on. We recommend that you
|
|
define
|
|
the list with odd values simply for consistency. <br>
|
|
</p>
|
|
<h4><a name="ContourSubSample"></a><font size="+1">Contour Sub
|
|
Sample</font></h4>
|
|
The Countour Analyzer algorithm is used to convert contours into a grid
|
|
for the <a href="GFETrainingSpatialEditor.html#ContourTool">Contour
|
|
Tool</a>.
|
|
It can run over a subsampled grid to improve performance. This is
|
|
usually ok since the contour tool is mostly used where there is not
|
|
much
|
|
detail due to topography. The value of ContourSubSample is used to
|
|
divide
|
|
the x and y dimensions of the original grid to get the dimensions of
|
|
the
|
|
subsampled grid. So, setting ContourSubSample to 4 would cause the
|
|
Contour
|
|
Analyzer to reduce a 400x400 grid to a 100x100 grid for contouring
|
|
purposes.
|
|
This can greatly speed up the algorithm. Setting ContourSubSample
|
|
to 1 will cause no reduction. The default value is 4. If
|
|
ContourSubSample
|
|
is set to a value less than or equal to 0 then it will go back to
|
|
4.
|
|
If it is set to a value large enough to make the subsampled grid have
|
|
an
|
|
x or y dimension less than 5, then it will be reduced so that the
|
|
minimum
|
|
dimension for x or y will be 5. This configurable item is simply
|
|
for performance.
|
|
<p><b><tt>ContourSubSample = 4</tt></b> <br>
|
|
</p>
|
|
<hr style="width: 100%; height: 2px;">
|
|
<div class="Body">
|
|
<center><a href="GFESuite.html">Back To TOC</a></center>
|
|
</div>
|
|
</div>
|
|
</body>
|
|
</html>
|