Climatogical historical data question

[DougW]

I don't deal with climate data much so I have a question on how to use this data. I've been given Mean- Mean , Mean-Min, Mean -MAx, extreme min and extreme max for each month and the year overall. I was thinking that the mean min and max were std deviation but not same amount off the mean heres data for overall year

mean mean 28.5 °C
mean max 36.6
mean min 20.3
extreme max 48.2
extreme min 7

and yes its all in °C and its a hot place

also give means of extreme min and maxs though can't see the point of that

Would like if I could find ~ running average low for 6 hours or something like that. The min temperature I really don't care about its whats the average temp a period of time.

Also have daily min max mean for a Jan, so I know that the min temp of 7° day had a mean of 17 and a max of 24. Could you worst day data and assume temp ... and fudge a temp distribution that works out to the mean?

[xboat]

i nied pikchers...i kant reed so gewd.

[L7]

That post is mean.

[grrrr]

so you have the mean min max and the max min mean, but the mean mean is mean and you need a fudge sundae?

[L7]

Less smart ass now. I have seen data for some areas with mean hourly temperatures that may be more realistic. Not sure what you want from this but I always took mean to be the median between high and low regardless how long it may have been at any temp. So if your high is 32 but 8 hours of the day were between 30 and 32 but a low of 16 was hit but it only spent 1 hour within 2 degrees of that low your mean would still be 24 which would be deceptively cool given what averaging the temp from each hour would give you.

Is the mean actually the hourly average?

[runethechamp]

I'm a nerd so I can help you if you have some more data. Looks like you need the raw data to figure out what you need.

[runethechamp]

Less smart ass now. I have seen data for some areas with mean hourly temperatures that may be more realistic. Not sure what you want from this but I always took mean to be the median between high and low regardless how long it may have been at any temp. So if your high is 32 but 8 hours of the day were between 30 and 32 but a low of 16 was hit but it only spent 1 hour within 2 degrees of that low your mean would still be 24 which would be deceptively cool given what averaging the temp from each hour would give you.

Is the mean actually the hourly average?

Well, taking the mean as the average of max and min isn't usual, is it? Depending on sample size, mean and median can be very different.

[Graupel]

I don't review climatological historical data often, so I'm not sure if there is a standard definition for these values. However, this is what I'm guessing. I could be totally off.

note: just for clarity, the mean is the same as "average" which is the sum of the values in which you are interested, divided by the number of values you have.

I agree with Champ, that the mean (daily at least) probably isn't the average between the min and the max. It's probably a time weighted average as L7 suggests.

Let's say this data set has a min temperature, a mean temperature, and a max temperature for every day in a month. For a 31 day month, that's 31 mins, 31 means, and 31 max's.

The "mean min" for the month would be the average of the 31 min temperatures.
The "mean max" would be the average of the 31 max temps.
The "mean mean" would be the average mean temps.

The "extremes" would just be the absolute lowest temp recorded that month: "extreme min", and the highest temp for the "extreme max"

Does this seem to be the case with your data set?

If not, let me know. I'll drink a beer and look at the question again. That always helps.

[Telenater]

I'm a nerd so I can help you if you have some more data. Looks like you need the raw data to figure out what you need.

I think that you're correct, but I'm only 95.44997% sure.

[DougW]

I'm a nerd so I can help you if you have some more data. Looks like you need the raw data to figure out what you need.

Well in a way rather not have raw data as would take too much time.

L7 - I'm sure the mean is the average but I was thrown off by what the min mean was. But did a search and it is a term used by weather folks though can't find a definition yet.

[runethechamp]

L7 - I'm sure the mean is the average but I was thrown off by what the min mean was. But did a search and it is a term used by weather folks though can't find a definition yet.

I think Graupel is correct in his assumptions. At least within two standard deviations

[Telenater]

Without knowing more about the data.

I'd guess
Mean = average temperature for the month based on temperatures taken at equal intervals.

Mean Max = Average of the daily high temperatures for the month.

Mean Min = Average of daily low temperatures for the month.

max max= maximum temp recorded during the month
Min min= minimum temp recorded during the month.

edit: What Graupel said.

[DougW]

I don't review climatological historical data often, so I'm not sure if there is a standard definition for these values. However, this is what I'm guessing. I could be totally off.

note: just for clarity, the mean is the same as "average" which is the sum of the values in which you are interested, divided by the number of values you have.

I agree with Champ, that the mean (daily at least) probably isn't the average between the min and the max. It's probably a time weighted average as L7 suggests.

Let's say this data set has a min temperature, a mean temperature, and a max temperature for every day in a month. For a 31 day month, that's 31 mins, 31 means, and 31 max's.

The "mean min" for the month would be the average of the 31 min temperatures.
The "mean max" would be the average of the 31 max temps.
The "mean mean" would be the average mean temps.

The "extremes" would just be the absolute lowest temp recorded that month: "extreme min", and the highest temp for the "extreme max"

Does this seem to be the case with your data set?

If not, let me know. I'll drink a beer and look at the question again. That always helps.

Was taking so long to reply I didn't see yours. Yes that makes sence. min mean being the average low temp recorded over a month. Not too useful a number for me.

I guess I need and average day correlation with input mean temp, high and low that would give me the average temp each hour- I'll use a sin see if that works.

[Graupel]

If you have 3 data points a day: min, mean, and max, then I don't think you have enough data to create a reasonable distribution for a 6 hour running average. I just don't think you could really have any confidence in the results.

I'm only part time nerd though, so I may be wrong.

Why do you want a 6 hour running average anyway?

[runethechamp]

If you have 3 data points a day: min, mean, and max, then I don't think you have enough data to create a reasonable distribution for a 6 hour running average. I just don't think you could really have any confidence in the results.

I'm only part time nerd though, so I may be wrong.

Why do you want a 6 hour running average anyway?

I don't know enough about this to be sure here, but I thing a sine/cosine function could be a reasonable approximation for temprature variation over 24 hrs. But, I don't know how easy it is to fit a curve to these three points if the max and min is something else than 12 hours apart. Also remember that the ends of the curve for each day should match the ends of the curves of the previous and next day.

[DougW]

I guess making assumption that temp distribution is sine shape after that fits pretty well as lucky on the day as mean is 14, max is 21 and low is 7 . Just find the area under curve to get average temp for 12 hours. Works there as day is about 12 hours long all year as close to equator. Reason for 12 hours as have water tank to feed system sized for 12 hours, will size for highest demand... will have to thnk about that.. Better than client was going to do as going to size for lowest temp though only occurs for a couple of hours.

[L7]

Well, taking the mean as the average of max and min isn't usual, is it?

Well I don't know so that's why I asked.

Doesn't this job cry out for Red Baron, this is his gig isn't it? I assume he would have these definitions down pat.

[jdabasin]

http://www.crh.noaa.gov/cgi-bin-den/showProduct.pl?title=Denver's+Climatological+Summa ry<br>Current+Month+to+Date&product=wrkf6&backto=2


the daily mean is just the average of the daily min and the daily max per National Weather Service standards. Check it on this summary for Denver data. They don't typically bother taken hourly values and averaging those over 24 hours as data points are missed. And in the long-term, averaging 24 hours of hourly temp. data to get daily mean temps versus using just the avg. of the max and the min produce nearly identical results. Maybe not on a case by case basis though. wow. stats are dumb.