To our amazement, the previous analysis showed that based on temperature lows and highs, the Charleston site very closely followed that of the Observatory. With this information, we then contacted the MIC at the Charleston branch of the National Weather Service to find out if they had archived RAOB soundings at their disposal. Then once again to our surprise, he had told first that they have not had any RAOB soundings in quite some time, and also, he had told us not to look at the weather of Charleston, but rather that of Pittsburgh, which is to the north, and where the predominant weather is coming from. This itself made sense to us, but also, being that Charleston is horizontal of the Observatory, that its vertical atmosphere would be similar, and similarly changed to that of the site of NRAO. Unfortunately, the only vertical data we had available to us for Charleston were a few GOES soundings, and Bufkit profiles which were still unknown how accurate they were.

To discover this, we decided to compare each type of sounding to that of the READY sounding, which is directly overhead of the observatory. This analysis would prove to show which type of vertical profile of the atmosphere would be ideal to use as a model for the Green Bank site. This showed that within RAOB soundings, it was more ideal to look at the Pittsburgh RAOB sounding rather then that of the Blacksburg RAOB. We decided this due to the Pittsburgh sounding better following the same trends in absolute and dew point temperature through rising pressure levels. We also compared GOES soundings to the general shape, and temperatures at varying pressure levels to that of a READY sounding. This analysis showed that the Charleston sounding more closely followed that of the READY sounding in both absolute temperatures, and that of dew point. The final analysis between vertical sounding models, and instruments was the Bufkit forecast model compared to that of a READY diagram. This showed that the sounding station to the north of the NRAO site didn't follow the READY diagram very well. This site was the Pierce site. There were several other sites analyzed and compared, and two were both very close to in both temperatures, and general shapes. These sites were the Roanoke sounding site, and the Charleston sounding site. It was difficult to discern which was more accurate in the comparison.

The next set of data we had to analyze and sort in order to determine what times specifically were ideal for observation with the Green Bank Telescope. This is able to be done was the Tipper, and 100 meter Interferometer. The tipper is an 86 GHz, radiometer that operates at a center frequency of 86 GHz (about 16.5 millimeter wavelength). This device scans at 360 degrees every second, and takes its measurements for every one minute, then integrates it. The tipper measures Opacity, which is a function of the line of Tau. More simply put, the tipper measures how good the seeing is at a given moment. The 100-meter Interferometer is a pair of small radio telescopes 100 meters apart so as to match the diameter of the GBT (Green Bank Telescope). What this measures is when a radio signal from a geocentric satellite is in phase with itself at both telescopes, when it matches.

What the next step from this is to determine what is ideal for observing at high frequencies. This was previously determined during our teacher work during the summer. This was when Tau is under .1, and Phase is under 300 microns. Also, to be sure that the same weather conditions are not present during poor observing times, a list was compiled initially at a minimum of twice as bad as what is considered good observing. This would be at least .2 Tau, and at least 600-micron variance in Phase. The problem with this became that there were to few data sets in comparison to the good observing data sets, so we had changed poor observing minimum requirements to be no less then 1.2 Tau, and no less then 400 micron variance in Phase.

This data was collected by NRAO, and then sent to us via an FTP server. We then used plotting, and data analysis software to reduce data, and then combine data. In order to do this, a TCL program was written by Ron Madelena to as closely as possible match dates within a data file with that of another. But before we did this, we had to reduce the data to the two possibilities we were looking for. This would be good observing times in opacity, then, later, bad observing times from the same initial data file which was not reduced. Next, the reduction was done to find both good and poor observing times in the 100-meter interferometer. The next phase in this process was combine these the two categories of files, good and poor observing into two master files, one showing both good observing in opacity, and in phase. The other would be a combination of what was poor observing in opacity, and poor observing in phase. Once the TCL program completed, it matched the dates as closely with what was given between the two input files, then, in order to see which dates where they were further apart, a column was created to show variance. Due to the sensitivity of the research, we needed a variance within matched dates of no greater then .007 days, which is just over 10 minutes in length.

Following the previous step, it becomes necessary to then make a spreadsheet to liken the weather data collected to the corresponding times and dates for both good, and poor observing. With this in mind, and based upon the tried knowledge of the MIC in Charleston, we collected RAOB data from both Pittsburgh to the north, and Blacksburg to the south. Next, based upon our hypothesis, decided what data given from the data found on the RAOB to get. This included looking at the height of the tropopause, how much precipitable water there is in the atmosphere, relative humidity of the atmosphere, the thunderstorm K Index to see what kind of movement there exists in the atmosphere, if there were any wind shears, and to what skewed direction off kilter theses shears were.

Also in correspondence with the good, and poor observing time periods, we the input data into two categories, one with the vertical atmospheric data found during good observing times, and the vertical atmospheric data found during poor observing times. Once this data was input, then within each column, it was averaged, and a range from the least to the greatest numeral was found. What this would show is if there could be any trends within the data, we could see an average trend, which is depicted in good, or poor observing. Also, we could see how accurate a trend could be.

This type of analysis was also done with the READY diagrams. Except, with the READY model there are no pre-calculated variables displayed. So instead of using the given calculations like used in the RAOB soundings, we decided to input the data of the absolute temperature, Ta, and the temperature of dew point, Td at four pressure levels. These levels started at the ground level, then to the 850mb level, next was the 600mb level, and finally, we had taken temperature readings off the graph at the 600mb pressure level.

There was an additional set of procedures we partook in to better see if there were any hidden trends. These additional procedures included taking Td of any higher level subtracted from the Td of the ground level of the proceeding lower level where the ground layer is the lowest. This procedure was also followed with the absolute temperature, Ta. The final analysis of the READY diagrams was taking the absolute temperature at a given level and subtracting the temperature of dew point.