PATMOS-x Comparison to Other Climatologies

Linear Correlation

Anomaly Magnitude

Anomaly Correlation with PATMOS-x

Anomaly Correlation with ISCCP

NOAA+UW HIRS

NCEP Reanalysis

ERA40 Reanalysis

PATMOS-x Comparisons with other Satellite Cloud Climatologies

This page shows some results prepared for the Second GEWEX Cloud Climatology Assessment Working Group Meeting to held in July, 2006 in Madison, Wisconsin. These results are preliminary and are likely to change. We have limited the analysis to the months of January, April, July October because that is only data available from the latest AVHRR Pathfinder Atmospheres Extended (PATMOS-x) processing. All data from PATMOS-x should be available by the time of the meeting. The analysis codes are written in IDL and available upon request.

The total cloud amounts where taken from the PATMOS-x field named "frac_total_cld_ir" and the high cloud amounts where taken from "frac_high_cld". For ISCCP-D2 data, the total cloud amount was taken from record #8 and the high cloud amount was taken from record #38. (We will update the ISCCP high cloud amounts to include vis/ir results). For NOAA+UW HIRS, the high cloud amount was taken as the fourth reported level and the total cloud amount was taken as the tenth reported level. The total cloud amount from NCEP was computed as the mean from the 6 hourly values from each day.

More PATMOS-x analysis (EOF and diurnal cycles) can be viewed by Amato Evan's page and other non-PATMOS-x results for this meeting can be accessed via the main website.

Results on this page

Correlation between local and global anomalies
Spatial Patterns of Linear Trends and Anomaly Correlations Between Data-sets
Time Series of Cloud Amounts
Comparison of PATMOS-x monthly cloud amounts with those derived from the ARM CART MMCR
Comparison of PATMOS-x and MODIS collect 5

Correlation Between Local and Global Anomalies

The following images show the correlation between the anomalies for each grid-cell and the anomalies in the global means. The anomalies are computed relative to a linear fit through each time series (global or grid-cell). Regions of high correlation indicate regions where the local variation is contributing to the global variation. In this analysis, the PATMOS-x data is at it's native resolution of 0.5 degree and HIRS is at it's native resolution of 1.0 degree.

Total Cloud Amount

	January	April	July	October
PATMOS-x 1983 - 2004
ISCCP D2 1983 - 2004
NOAA+UW HIRS 1985 - 2001
NCEP Reanalysis 1983 - 2004
ERA40 Reanalysis 1983 - 2001

High Cloud Amount

	January	April	July	October
PATMOS-x 1983 - 2004
ISCCP D2 1983 - 2004
NOAA+UW HIRS 1985 - 1999
ERA40 Reanalysis 1983 - 2001

Spatial Patterns of Linear Trends and Anomaly Correlations Between Data-sets

These results compare the climatologies to each other. The time period (1985-2001) was the common period among all data-sets. All data (when needed) was converted to 2.5 degree equal angle grids for analysis. Daily averaged fields were used.

A brief description of the statistics shown below

Mean - the mean of the time series in each cell
Trend - the slope of a linear fit to the whole time series in each cell. Cells where the linear correlation was less than 0.25 are shown as missing (gray).
Linear Correlation - The correlation of a cell's time series with the linear fit to that time series.
Anomaly Magnitude - Mean absolute value of the anomaly for each cell
Anomaly Correlation with PATMOS-x - The correlation of the anomaly time series between PATMOS-x and another climatology for each cell. A value of 1.0 means the data are strongly correlated, 0 is no correlation, -1 is strongly anti-correlated

Total Cloud Amount for January

Total Cloud Amount for April

Total Cloud Amount for July

Total Cloud Amount for October

High Cloud Amount for January

High Cloud Amount for April

High Cloud Amount for July

High Cloud Amount for October

Time Series of Cloud Amounts

These results compare the zonallly averaged cloud amount time series to each other. The time period (1985-2001) was the common period among all data-sets. All data (when needed) was converted to 2.5 degree equal angle grids for analysis. Daily averaged fields were used. There is no NCEP High Cloud Product within the NCEP reanalysis output available to us.

	January	April	July	October
Total Cloud Amount 60S to 60N
Total Cloud Amount 20S to 20N
High Cloud Amount 60S to 60N
High Cloud Amount 20S to 20N

Comparison of PATMOS-x monthly cloud amounts with those derived from the ARM CART MMCR*

Monthly Mean of Daily Averaged Total Cloud Amount
Monthly Mean of Daily Averaged High Cloud Amount
Ratio of Total to High

* Data provided by Jay Mace from the University of Utah. High cloud from MMCR defined as cloud above 450 mb. PATMOS-x defines high cloud as cloud with cloud-tops above 440 mb.