TOC  |      |      |              |              

stree

The stree program is used to pick the baselines for a session. The program does this by finding the "minimum spanning tree" that connects all the stations. The user specifies which stations to use as "hubs" (if any), which stations are unreliable and thus must be connected by a "spur" baseline, and which baselines are to be forced to be part of the spanning tree by the user. The input files for these choices are called hubsite.inp, spursite.inp, and fixline.inp, respectively. The minimum spanning tree that is chosen will always be a set of independent vectors that connects all the stations in the session. If the user wants to add a redundant, "dependent" vector then that can be done after the stree program (by adding the extra baseline to the aazz00.bls file).

The stree program reads in a control file called stree.inp which has the following five lines:

The stree program decides the number of common satellite observations between each pair of stations by reading in the elv.plt file. It determines the baseline lengths between each pair of stations by reading in the *hd.dat file from the multi-station mergedb run. Let's say the user were to specify hubsites, spur sites, and fixed baselines for a session, then the stree program (as a maximum) would output the following seven sets of independent baselines to the stree.sum file:

1. The baselines chosen using the baseline lengths only.
2. The baselines chosen based on the maximum amount of data only. (Note: the weight matrix for the amount of data starts with the "compts" (common points) matrix. This compts matrix has the #epochs common to BOTH stations stored in the upper triangular part, and the #epochs observed by a least ONE station in the lower triangular part. Small numbers in the upper part may indicate gaps in the data for some stations.)
3. The baselines chosen using some combination of "baseline length" and "maximum data"(For example, the stree.inp file might specify to weight the baselines 80% by length and 20% by the amount of data).
4. The baselines chosen using the (80/20) percentage weights, AND using only baselines that are connected to a hub site.
5. The baselines chosen using: the (80/20) percentage weights, the baselines connected to a hub site, AND forcing all spur sites to be connected by a "spur".
6. The baselines chosen using: the (80/20) weights, the hub sites, the spur sites, AND the fixed baselines listed in the fixline.inp file.
7. All the baselines listed above in (6) but sorted by hub site and longitude (this is done to try to make the covariance matrix for the double difference observables in pages as close to a block-diagonal matrix as possible).

If the percentage weights are set at 100.0 and 0.0, then the unconstrained baselines will be chosen based only on length. If the weights are set at 0.0 and 100.0, then these baselines will be chosen based only on the amount of data for the baselines. To decide how best to set these weights, the user can run stree one time for the session and look at the first two sets of baselines. If they are significantly different from each other, then it may mean that some baselines have numerous missing epochs; thus the user might want to increase the weight that is based on the amount of data.

The stree program sorts the final set of baselines by hub site and by longitude. All baselines emanating from the same hub site are grouped together. The groups are sorted from west to east based on the longitude of the hub sites. Then within each group, the baselines are sorted from west to east based on the east longitude of the non-reference station. The final sorted baselines are output to the aazz00.bls file along with their 4- char database prefix (e.g.,acab, adab, etc.)

The stree.sum file shows all the input files found by the program and all the various set of baselines as mentioned above. The top of the stree.sum files also lists the satellites found in the *hd.dat file; a list of latitude, longitude, and height for each station; and the baseline lengths for all possible combinations of stations. The stree.sum file will also list error messages whenever a critical input file is missing. The critical input files are: stree.inp, aazz00hd.dat, and elv.plt.

The input and output files used in stree are summarized below:

Input Files:
-----------
stree.inp
spursite.inp
hubsite.inp
fixline.inp
aazz00hd.dat
elv.plt

Output Files:
------------
stree.sum
aazz00.bls

 TOC  |      |      |              |