Help interpreting NOAA weather data, please
February 14, 2016 11:55 AM Subscribe
I'm looking at this NOAA page: U.S. Percentage Areas (Very Warm/Cold, Very Wet/Dry) and I'm having trouble understanding what they are representing. Thank you in advance to meteorologists/statisticians/anyone who can shed light on this!
The link on the main page says:
U.S. Percentage Areas (Very Warm/Cold, Very Wet/Dry)
This product shows, for each month, season and year since 1895, how much of the contiguous United States experienced extreme (top or bottom 10%) temperature and precipitation conditions.
The explanatory paragraph on the page with the data says:
The percentage areas of the contiguous United States are computed based on the U.S. Climate Divisional Dataset. Those climate divisions having the monthly average temperature/total precipitation in the top ten percent (> 90th percentile) of their historical distribution are very warm/wet and those in the bottom ten percent (< 10th percentile) are very cold/dry.
The first data point (and I've set the month to August) shows:
August......Very Warm..........Very Cold
1895...........1.90%.................3.50%
It looks to me like they are saying "In August 1895, 1.90% of the contiguous US was the warmest and wettest of all temperatures/precipitation recorded that year." I.e., if the range of recorded temperatures for all of the US in August 1895 was 0F-100F, 1.90% of the US had temperatures of 90-100F?
I.e., if the world is getting warmer, we should be seeing an increase in that statistic as the years go on?
If it matters, I have to do a simple regression analysis project (very basic and limited in scope for a survey-type course) and my equation is testing if there is any correlation between production levels of a specific agricultural product and pesticide use and/or climate change. So my equation would be something like:
Y = f(+Production, -Pesticide, +AugustWarmWet, -JanuaryColdDry)
Production: Pounds of Agricultural Product in the US per year
Pesticide: Pounds of Pesticide Applied in the US per year
AugustDry: Percentage of the US that was extremely hot and wet in August on the specified year
JanuaryWet: Percentage of the US that was extremely cold and dry in January of the specified year
Thank you!
The link on the main page says:
U.S. Percentage Areas (Very Warm/Cold, Very Wet/Dry)
This product shows, for each month, season and year since 1895, how much of the contiguous United States experienced extreme (top or bottom 10%) temperature and precipitation conditions.
The explanatory paragraph on the page with the data says:
The percentage areas of the contiguous United States are computed based on the U.S. Climate Divisional Dataset. Those climate divisions having the monthly average temperature/total precipitation in the top ten percent (> 90th percentile) of their historical distribution are very warm/wet and those in the bottom ten percent (< 10th percentile) are very cold/dry.
The first data point (and I've set the month to August) shows:
August......Very Warm..........Very Cold
1895...........1.90%.................3.50%
It looks to me like they are saying "In August 1895, 1.90% of the contiguous US was the warmest and wettest of all temperatures/precipitation recorded that year." I.e., if the range of recorded temperatures for all of the US in August 1895 was 0F-100F, 1.90% of the US had temperatures of 90-100F?
I.e., if the world is getting warmer, we should be seeing an increase in that statistic as the years go on?
If it matters, I have to do a simple regression analysis project (very basic and limited in scope for a survey-type course) and my equation is testing if there is any correlation between production levels of a specific agricultural product and pesticide use and/or climate change. So my equation would be something like:
Y = f(+Production, -Pesticide, +AugustWarmWet, -JanuaryColdDry)
Production: Pounds of Agricultural Product in the US per year
Pesticide: Pounds of Pesticide Applied in the US per year
AugustDry: Percentage of the US that was extremely hot and wet in August on the specified year
JanuaryWet: Percentage of the US that was extremely cold and dry in January of the specified year
Thank you!
Best answer: They have divided the continental U.S. Into 344 climate regions. For each region, the monthly average temperature is compared to the same region's long-term record. In your example, the regions where August 1895 was one of the 10 percent warmest of all Augusts in the period of record for that region added up to 1.9% of the total area of the continental U.S.
posted by jkent at 12:19 PM on February 14, 2016
posted by jkent at 12:19 PM on February 14, 2016
Best answer: It looks to me like they are saying "In August 1895, 1.90% of the contiguous US was the warmest and wettest of all temperatures/precipitation recorded that year." I.e., if the range of recorded temperatures for all of the US in August 1895 was 0F-100F, 1.90% of the US had temperatures of 90-100F?
What this means is that, for every area, we take the August 1895 warm/wet value and we compare it to the set of historical values for that area. If it is in the top 10% of those historical values (i.e. higher than 90% or more of the historical values), then we add that area to the total. In other words, in August 1895, 1.90% of the total area was in the top 10% of the historical values for that area. We are comparing the value for a particular area in August 1895 to the historical record for that same area. We are not comparing any particular area to the range in the US.
If the world is getting warmer, you may expect to see some increase in the warm/wet values, but this is not a simple correspondence.
If you want to test something against climate change, this is definitely not the data you should be testing against. Perhaps global average annual temperature would be appropriate? You may also want to look at C02 emissions or levels in the atmosphere as climate change lags CO2 emissions significantly.
posted by ssg at 12:30 PM on February 14, 2016
What this means is that, for every area, we take the August 1895 warm/wet value and we compare it to the set of historical values for that area. If it is in the top 10% of those historical values (i.e. higher than 90% or more of the historical values), then we add that area to the total. In other words, in August 1895, 1.90% of the total area was in the top 10% of the historical values for that area. We are comparing the value for a particular area in August 1895 to the historical record for that same area. We are not comparing any particular area to the range in the US.
If the world is getting warmer, you may expect to see some increase in the warm/wet values, but this is not a simple correspondence.
If you want to test something against climate change, this is definitely not the data you should be testing against. Perhaps global average annual temperature would be appropriate? You may also want to look at C02 emissions or levels in the atmosphere as climate change lags CO2 emissions significantly.
posted by ssg at 12:30 PM on February 14, 2016
Response by poster: Thanks to both of you for your help. I think, after reading your interpretations, I'll have to find different temperature/drought data. ssg, the CO2 emissions thing is interesting and not something I'd considered.
Cheers!
posted by Beti at 9:40 AM on February 16, 2016
Cheers!
posted by Beti at 9:40 AM on February 16, 2016
This thread is closed to new comments.
posted by Beti at 11:57 AM on February 14, 2016