Throughout 2001, the delivered seeing was regularly monitored with UFTI. Whenever UFTI is in use a routine automatically performs photometry on all frames taken with UFTI and fits a Gaussian to images in order to calculate the FWHM. The FWHM are output to a text file which is then read by a code and the statistics analysed. In order to do this we only look at faint standards, as we know these are all point sources. The code also corrects for airmass (i.e. corrected to the zenith equivalent where airmass is 1.0), and for wavelength (i.e. it is corrected to the K band equivalent).

The UT coverage is binned into 5 bins, each covering 2 hours in UT from 6 to 16 hours. The seeing is binned into 20 bins with the total coverage starting at 0″ and ending at 2″. However, it should be noted that anything below 0.”2 is very questionable and is most likely due to insufficient S/N. The resulting histograms are shown in figure 1 as well as a plot of seeing versus UT time of night. Table 1 also shows the median seeing for each UT bin and 2 extra UT bins at the beginning and end of night. The additional bins at UT 4-6 hours and UT 16-18 hours, both have significantly lower numbers than the other UT bins and so their histograms are not shown in figure 1. Both figure 1 and table 1 demonstrate how seeing varies through the night. The histogram for the UT range 6-8 hours in figure 1 shows that by this time, the atmosphere is starting to stabilize. As the night progresses and the telescope structure cools the seeing gradually improves. This results in the best seeing of the night being observed after 2am HST. The histogram representing the UT range 12-14 hours demonstrates this clearly, with the seeing statistics quickly rising to 0.”5 and then tailing off more slowly. The median seeing for this time of the night drops to approximately 0.”53 (see table 1). The overall median seeing for 2001 was 0.”6 

UT range	Median seeing
04-06 hrs	0.”79
06-08 hrs	0.”65
08-10 hrs	0.”62
10-12 hrs	0.”59
12-14 hrs	0.”53
14-16 hrs	0.”54
16-18 hrs	0.”69

Table 1: This table shows how median seeing varies with time of night.

• 

• 

• 

• 

• 

• 

Figure 1: Seeing versus time of the night and histograms showing how seeing varies as a function of the UT time of the night. 

Table 2 shows seasonal variations in the seeing statistics. Each number in the table represnts how often (as a percentage) in a given month, seeing is below a given value. August stands out as the best month, with seeing being better than 0.”5 arcsec 49% of the time. July and September follow August as the next best months. November seems to be something of an anomaly. UFTI was rarely used during November, and then only towards the end of the month as we were hindered by bad weather and non-photometric conditions. As a result UFTI was only in use during exceptional conditions. Early December was also affected by a storm for a few nights, but after this cleared the rest of December was observable.

Seeing

Month	0.”3	0.”4	0.”5	0.”6	0.”7	Number of  data points
January	0.48	3.8	19.8	41.8	61.6	(627)
February	0.52	8.8	23.5	44.6	64.5	(774)
March	0.80	1.2	6.8	20.1	45.6	(502)
April	0.42	9.6	32.7	62.7	78.4	(1185)
May	1.09	9.0	34.0	59.2	74.0	(736)
June	0.00	5.3	20.2	37.2	54.3	(94)
July	1.71	15.2	32.3	57.8	75.8	(1228)
August	2.84	27.3	49.0	64.1	79.7	(1269)
September	0.29	11.7	40.8	63.3	81.0	(343)
October	1.09	9.4	20.7	29.7	36.6	(276)
November	6.66	45.0	70.8	92.5	99.2	(120)
December	0.00	2.2	11.2	27.7	46.0	(365)
All	1.24	12.6	31.1	52.6	69.8	(7519)

Table 2: Table showing how often seeing is better than a particular value, i.e. in Jan seeing is better than 0.3 arcsec 0.48% of the time, better than 0.4 arcsec 3.8% of the time, better than 0.5 arcsec 19.8% of the time, etc. The number in brackets represents the number of data points available for each month.