NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Dynamic > KeyIssues > DaytimeNighttimeLSTRange

Changes | Index | Search | Go

Daytime, Nighttime and LST Range

• The effect of thermal gradients on antenna surface efficiency, and why we bother with "daytime" and "nighttime"
• The definition of "daytime" and "nighttime"
• The definition of LST Range

The effect of thermal gradients on antenna surface efficiency, and why we bother with "daytime" and "nighttime"

One of the main environmental factors which degrades telescope performance are thermal gradients in the antenna structure, which cause the antenna to distort from its "ideal" (or at least its "thermally neutral") shape. The effects on pointing/focus are at least partially compensated for by the thermal monitoring system and associated "dynamic pointing corrections". In the case where this correction is not enough, pointing/focus can be brought into acceptable limits, day or night, by sufficiently frequent local pointing/focus corrections. We can therefore effectively compensate for the effects of thermal gradients on pointing/focus.

In principle, the effects of thermal gradients on mirror figure and alignment can also be corrected (e.g. by out-of-focus holography). In practice, these observations currently take too long for practical real-time use. As a result, the antenna surface efficiency, and hence overall observing efficiency, can degrade to unacceptably poor levels in the presence of large thermal gradients - i.e. during "daytime". The loss of surface efficiency due to a given amount of mirror deformations/misalignment - characterized as the "equivalent root mean square surface error" - goes approximately as the exponential of the square of the observing frequency - hence the efficiency falls off extremely rapidly as the observing frequency increases. (I should put in some quantified numbers here).

Other environmental effects (e.g. high winds, poor atmospheric transmission) can also degrade observing efficiency. However, many of these effects are only weakly correlated with time of day, and are also difficult to predict more than a few hours/days in advance. By contrast, we know that "daytime" thermal gradients will degrade observing efficiency - and we can predict "daytime" arbitrarily far in advance. Therefore, instead of "dynamically" scheduling around thermal conditions, we define the following rules well in advance:

• Q-band (40 GHz) and above observations should never be scheduled during "daytime".
• "By default, we assume that daytime observing will be acceptable for all observations below 26 GHz [i.e. K-band and below]. and that all observations above 26 GHz [i.e. Ka-band and above] will occur at night. Observers with proposals below 26 GHz must provide full justification in their proposals if they require nighttime observing." (quoted from The Performance of GBT Receivers).
• Observations at Ku-band (18 GHz) and below are not seriously affected from the perspective of antenna surface efficiency by thermal gradients.

In principle, one might ask why we don't dynamically schedule for "cloudy" daytimes, when the thermal gradients will be less severe. The most pragmatic reason is that if the conditions are "too cloudy" to build up significant gradients, the weather is too poor for very high frequency (> 40 GHz) observing in any event. This may not be true for 18 - 40 GHz observing, but if our process becomes so sophisticated that we want to accomodate this eventuality, everything else will be working pretty well!

The definition of "daytime" and "nighttime"

Thermal gradients are worst during the day, when the sun is actively heating the structure. The antenna takes some hours to return (closer) to equilibrium even after the sun has set. In the winter, it takes a few hours for the weak sun to begin to affect the structure. Accordingly, we currently define "daytime" to be two hours after sunrise until three hours after sunset. "nighttime" is simply the rest of the 24 hour period. We may adjust the boundaries of daytime and nighttime (and eventually even remove the daytime restrictions) as/when PTCS capabilities improve.

The range of daytime and nighttime are shown in the following figure, copied from the The Performance of GBT Receivers page, which has some further discussion related to this topic.

The definition of LST Range

The following could definitely be much better worded, but it is a start.

As part of scheduling Large Proposals, we intend to adopt a policy of allowing an individual proposal to use no more than a given presentage of the "available observing time in any LST range during any trimester" (see the Large Proposal Policy. What is the available observing time in any LST range?

It is quite possible that the formal definition already exists. I offer the following for the GBT, which may or may not agree with existing policy.

• We should explictly note we need "available nighttime observing time in any LST range" and "available daytime"... we want to reserve the nighttime LST time for high frequency only.
• Consider the "LST ranges" to be the 24 sidereal hours in the sidereal day (e.g. 0-1hrs LST, 1-2hrs LST, etc).
• Consider the plot above. On any given day, consider an LST range to be in "nighttime" if more than 50% of the hour falls within the grey region, otherwise it is in "daytime".
• Then for each LST range, simply count up the number of "days" (24 hour periods) within a trimester (four months) that it occurs as either nighttime or daytime. That is the total available nighttime (daytime) observing time for that specific LST range.

Consider the 0nC trimester (October - January inclusive). From the above plot, it is apparent that 8hrs has just snuck in to the end of nightime around 1st October. LST times move earlier with respect to civil time with each successive month; by January 31st, 8hrs LST corresponds approximately to midnight local time. Thus there are 123 hours (four months worth) of 8hrs LST "nighttime" in the 0nC trimesters.

By contrast, 20 hrs LST has moved into daytime by 1st October, and stays there the whole trimester. The only time 20hrs makes it to nighttime in "high frequency observing season" (October - May) is for the month of April and May. 20hrs LST thus has no available "nighttime" observing time in the 0nC trimesters, and only about 20 hours in the 0nA trimesters. This is one why it is in such high demand, as shown on the LST Pressure plot available from the Large Proposal Policy page.

-- RichardPrestage - 10 Oct 2006