<html><body><div style="color:#000; background-color:#fff; font-family:HelveticaNeue, Helvetica Neue, Helvetica, Arial, Lucida Grande, sans-serif;font-size:12pt"><div>Hi,<br><br>I'm working with some weather model (WRF) output and I'll like to<br>import the results into proj4-aware software (gdal or grass gis).<br><br>The LCC model domain specification takes center of region lat/lon as<br>input, not +lat_0, +lon_0. The values I've been given for +lat_1 and<br>+lat_2 are equal, and near to +lon_0. That doesn't look quite<br>right but the model seems happy enough with it:<br></div><div style="color: rgb(0, 0, 0); font-size: 15.6px; font-family: monospace; background-color: transparent; font-style: normal;"><br></div><div style="color: rgb(0, 0, 0); font-size: 18.7167px; font-family: monospace; background-color: transparent; font-style: normal;">the values I was given:<br></div><div><br>CEN_LAT = 40.26405f ;</div><div>CEN_LON = -76.99222f ;</div>
<div>TRULAT1 = 39.338f ;</div><div>TRUELAT2 = 39.338f ;</div><div>MOAD_CEN_LAT = 39.33799f ;</div><div>STAND_LON = -84.f ;</div><div><br></div><div style="color: rgb(0, 0, 0); font-size: 18.7167px; font-family: monospace; background-color: transparent; font-style: normal;"><br>What I do have of value is two of the raster grids put out by the<br>model which contain the lat and lon of each grid point, and these<br>seem correct. I also have two raster outputs for convergence angle,<br>ie the angle between grid north and geographic north at each grid<br>point. I know that +x_0 and +y_0 are 0, and that +ellps is a sphere,<br>the WRF code lists about 4 different radii for +a, but that's a<br>secondary issue. Grid cell size is down to 4km, so I'm not overly<br>concerned with datum issues on the order of 100m.<br><br><br>What I'd like to do is to take the center-of-region LCC definition<br>and solve for the +lon_0, +lat_0, +lat_1, +lat_2 terms so I
can<br>process the model's output grids in their native projection.<br><br>I could treat the two lat,lon grids as source data to run a manual raster<br>reprojection and interpolate into a proj4-know projection, but that's<br>computationally expensive and lossy.<br><br>So far I've gotten close by choosing the bottom-left grid point of<br>the lat and lon grids, and reading off those values to use for the<br>+lat_0 and +lon_0 terms. Then by educated-guessing I picked some<br>+lat_1 and +lat_2 terms and iteratively have gotten pretty close to<br>being able to overly a geographic grid on top of contour lines from<br>the lat and lon raster grids. Of course the terms I found are no<br>where near those I was provided with or that I can see in the model's<br>domain setup files.<br></div><div style="color: rgb(0, 0, 0); font-size: 22.45px; font-family: monospace; background-color: transparent; font-style: normal;"><br></div><div style="color: rgb(0, 0, 0);
font-size: 26.9333px; font-family: monospace; background-color: transparent; font-style: normal;">my guess:<br></div><div style="color: rgb(0, 0, 0); font-size: 26.9333px; font-family: monospace; background-color: transparent; font-style: normal;">-PROJ_INFO--------------------<wbr>-----------------------------<br>
name : Lambert Conformal Conic<br>
proj : <span class="">lcc</span><br>
a : 6371229.000<br>
es : 0.0<br>
f : 0.0<br>
lat_0 : 33.0624<br>
lat_1 : 20<br>
lat_2 : 45<br>
lon_0 : -87.0188<br>
x_0 : 0.0000<br>
y_0 : 0.0000<br>
-PROJ_UNITS-------------------</div><div style="color: rgb(0, 0, 0); font-size: 26.9333px; font-family: monospace; background-color: transparent; font-style: normal;"><wbr>-----------------------------<br>
unit : meter<br>
units : meters<br>
meters : 1.0<br>
</div><div style="color: rgb(0, 0, 0); font-size: 32.3167px; font-family: monospace; background-color: transparent; font-style: normal;"><br>This trial and error gets me to within about 5% (see attatched), but<br>I'd really like to do the job properly and calculate the exact terms.<br>I have access to the Fortran code used in the model to make the lat<br>and lon grids, could I use them to also derive the proj4 terms?<br><br><br>thanks for any ideas or advice on how to proceed.<br><br><br>Hamish</div><div style="color: rgb(0, 0, 0); font-size: 38.7667px; font-family: monospace; background-color: transparent; font-style: normal;"><br></div><div style="color: rgb(0, 0, 0); font-size: 46.5167px; font-family: monospace; background-color: transparent; font-style: normal;">ps- sorry for any crazy formatting, yahoo mail's web app is broken yet again and won't let me change to ascii or have any control over the formatting. grumble!<br
class=""></div></div></body></html>