Web-based Reanalyses Intercomparison Tools (WRIT)



New! Ocean WRIT Talk at the fall 2020 AGU meeting.

Product Descriptions

Atmospheric or Oceanic Monthly/Seasonal Plotting Page:
Allows users to make maps and vertical cross-sections from various reanalyses. Statistics include means, anomalies and climatologies. Users can also difference the various reanalyses for all three types of statistics (for any overlapping date in the reanalyses). Composites (averaging multiple dates) are also available.

Atmospheric or Oceanic Monthly Timeseries Extractor/Analyzer:
Extracts timeseries at specified latitude/longitude or latitude/longitude ranges and either plots the time-series, time-series differences, and scatter plots. Various statistics are returned in addition to numeric values of the time-series plotted. Ability to plot and compare climate index time-series such as the PNA or Niño3.4 or upload your own.

Atmospheric or Oceanic Monthly Correlation Plotter:
Calculates correlations (and regressions) of the reanalysis and observational dataset with supplied atmospheric/ocean indices or user uploaded time-series. Users can plot correlations at different lead/lags. They can also create vertical cross-section correlation plots.

Atmospheric Trajectory Generator:
Allows users to plot forward and backward air trajectories from different reanalyses (currently NCEP R1, NCEP R2, and 20CR, and ERA-Interim). Users can plot the trajectories of one or more levels on a single plot. The output is plotted on a map and is available as netCDF and as KMZ files suitable for Google Earth.

Atmospheric Distribution Analyzer:
Allows users to plot and compare the distributions of daily data during a season from different sources including different reanalyses (currently NCEP R1 and 20CR) and observed data. Users can see where data values have fallen historically and can see how different distribution assumptions impact the tails of the distribution.

Atmospheric Vertical Profiles:
Allows users to plot different vertical products from reanalyses including skew-T, variable/height, timeXheight, and vertical transects (daily data).

Atmospheric Hovmollers:
Allows users to make plots of daily and subdaily time by latitude or time by longitude for different reanalysis datasets. Statistics include means and anomalies.


  1. Ocean WRIT Talk at the fall 2020 AGU meeting.
  2. Ocean WRIT: preliminary PPT


  • Smith C. A., G. P. Compo and D. K. Hooper (November 2014): Web-Based Reanalysis Intercomparison Tools (WRIT) for Analysis and Comparison of Reanalyses and Other Datasets. Bull. Am. Meteorol. Soc., 95 (11), 1671-1678. doi:10.1175/BAMS-D-13-00192.1

Future WRIT Tool Enhancements

  • Standardized time-series
  • Lead/Lag time-series relationships
  • Spatial correlations comparing different reanalyses

Future Atmospheric WRIT Products

We are exploring the feasibility of various tools. The features we hope to have are:

  • Daily time scale composites atmospheric reanalyses.
  • Sub-Daily composite atmospheric reanalyses
  • Creation and comparison of index timeseries such as the PNA from the various atmospheric reanalyses

Future Ocean WRIT Products

  • Comparison of buoy and profile data


  • Reanalyses.org: A wiki for advancing reanalyses. Users can find information about reanalyses (atmosphere, ocean, land), how to access and plot data, reanalyses observations, diagnostics, and more. Members can post information and ask questions of the reanalysis community.
  • Caveats. Also see specific tool information.
  • Smith C. A., G. P. Compo and D. K. Hooper (November 2014): Web-Based Reanalysis Intercomparison Tools (WRIT) for Analysis and Comparison of Reanalyses and Other Datasets. Bull. Am. Meteorol. Soc., 95 (11), 1671-1678. doi:10.1175/BAMS-D-13-00192.1


There are two types of gridded datasets in the PSL WRIT tools pages. They are atmosphere and ocean reanalysis and gridded observations/analysis. Each web tool uses different sets of datasets depending on what it is plotting/analyzing, our storage, server, and memory resources. There is a link from each page to the datasets used. There are also climate and ocean time-series available for the correlation webpage and the time-series plotting pages. See Users are welcome to suggest new climate index time-series or gridded observational datasets and we will consider adding them.

Atmospheric WRIT Examples

20CR plot example
WRIT: Composites Plot
20CR plot example
WRIT: Time-series Plot
20CR plot example
Trajectory Plot
2CR plot example
WRIT: Correlation Plot
20CR plot example
Vertical Profiles Plot
2CR plot example
Hovmollers Plot

Ocean WRIT Examples

godas ssh 1983 map
WRIT: Composites Plot
20CR plot example
WRIT: Time-series Plot
2CR plot example
WRIT: Correlation Plot
GODAS plot example
WRIT: Map (Depth) Plot
ORAS5plot example
WRIT: Correlation using user generated Time-series Plot

Ocean WRIT How to Guide

Mapping Tool 
How do I....?
Create map of a single variable? 
Choose a dataset and then a variable. Some variables are available at different depths so for those, also select a level. Select the season (first month to last) and the year. By using the defualt options you will get a plot.
Change the plotting options for a map of a single variable? 
Keeping the dataset/variable options, users can change the region shown, the colortable used, the type of shading, contour labels, type of fill, and the contour range and interval.
Examine anomalies for latest climate normal (1981-2010 long term average)? 
To use standard climatology to create anomalies, select statistic "anomaly". For most datasets, the long term mean for 1981-2010 is available and will be subtracted out.
Examine anomalies using a custom year range for the long term mean? 
To use your own climatology period, select the same dataset for dataset 1 and dataset 2. For the 2nd dataset, choose a range of years that represents your climatology. For dataset 1, choose the date(s) you are looking at. Select "mean". The plot and data returned will be an anomaly relative to the 2nd dataset time period.
Compare variables from different decades? 
Choose a dataset and variable. Select the same dataset and variable for dataset 2. For dataset 1, select a range of years and the season. For dataset 2 select a different set of years. For example, 2008-2017 for dataset 1 and 1988-1997 for dataset 2. Select mean for statistic. Choose JRA55, latent heat. Also, slect reverse colorbar.
Examine a climate process such as ENSO? 
Users can enter non consecutive years of interest. For example, using the Oceanic Nino Index (ONI), one can extract the 4 strongest El Niños for January: 2016, 1998, 1983, 1999. Enter those years into the boxes. Select a variable (u current) and level (5m) and season (Feb-Feb). Note the last month of the season defintion corresponds to the year entered. Select Mean.
Compare a climate process with 2 phases (El Niño and La Niña)? 
El Niño can be compared to La Niña by extracting a second set of years from the Oceanic Nino Index ONI that are lowest: 1989,2007,1996,2000 and entering them into dataset 2 boxes for the same dataset. In this case NCEP GODAS are used for both datasets and the 0-300m heat content is the variable.
Plot latitude by depth over a longitude range? 
Select a dataset (SODA-JRA55) and a variable with levels (u current). Choose longitude/depth. Values will be averaged over latitudes (2N to 4N) and plotted for 100E to 300E. In this case, the long term mean is plotted.
Plot vertical longitude by depth over a latitude range? 
Select a dataset (SODA-JRA55) and a variable with levels (potential T). Choose latitude/depth. Values will be averaged over longitudes (200 to 200) and plotted over latitudes (30S-30N). In the example, the long term mean is plotted.
Correlation Tool 
How do I....?
Create a map of time-series correlated with variable? 
Examine the spatial pattern of sea surface height (SSH) that corresponds with the Arctic Oscillation (AO) in January. Select NASA ECCO V4r4. For region, select northern hemisphere
Look at index timeseries leading a variable? 
Check to see what the 3 month lead sea surface temperature (SST) correlation is with Atlantic Hurricane Accumulated Cyclone Energy (ACE) for 1920-2019. Choose HadISST for SST, Atlantic Hurricane ACE from time-series list, June-June SST, and index lags by 3 months.
Map a trend? 
Select a variable and dataset and determine the trend. For example, select 5m salinity for SODA3-JRA55. Select season Septemper-September. Select trend from timeseries list. Change countour interval/range to .0125 and -1.25 to .125.
Use your own times-eries with the correlation tool? 
Upload Washington State rainfall for climate division division 1 (Olympic coast). Correlate the simultaneous correlation with NCEP GODAS zonal momentum flux for January-January. Choose "Pacific Basin" region and "Brown to Blue" color table.
Plot a vertical cross-section correlation plot? 
Correlate Niño 3.4 (El Niño) with potential temperature over the tropical Pacific for SODA3-ERAI. Note the average over latitude (or longiutude) is calculated first and then a correlation is calculated.
Time-series Tool 
How do I....?
Extract a timeseries from a latitude/longiture extent of a dataset? 
Select a dataset (NOAA OI SST) and variable (Skin T/SST). Select a rectangular region (Nino 3: -5 to 5, 210 to 270). Select Anomaly.
Compare the same time-series from 2 datasets? 
Select two datasets (SODA3-JRA55, NCEP GODAS) and u current at 5m. Select Anomaly. Select Nino3 region (-5 to 5, 210E to 270E). Choose key in bottom right.
Plot a pregenerated timeseries? 
Select "Pregenerated Time-series" for dataset. Select "Battery Park Tidal gauge". Plot 1900 to 2020. Select X range 1900 to 2025.
Create a scatter plot? 
Compare precipitation and salinity over the tropical Pacific. Choose dataset 1 SODA3-JRA55 salinity, 5m. Choose dataset 2 GPCP V2.3 precipitation. Select scatterplot. Enter -4 to 4; 200E to 200E.
Extract a time-series for use in correlation or time-series page? 
Select Dataset NCEP GODAS. Select variable net heat flux. Enter area 15N; 300E. On page with plot, select link to "Data Table". Save page as text in browser. You can upload to PSL web-tools including the ocean WRIT time-series or correlation tools.
Use your own time-series (1 or both)? 
From the extract a time-series above, upload it to anonymous ftp following instructions. Select pregenerated time-series and then user upload. Type your title and then the file name ("/Public/incoming/timeseries/yourname"). Select "NH ice extent" from pregenerated timeseries. Select season as August to August.
Create a land or ocean only time-series?  
Select SLP and the JRA-55 dataset. Enter -30 to 30, 0 to 360. Select ocean only. Select Time-series as plot type
Perform Wavelet Analysis?  
Select Nino 3.4 from pregenerated list of time-series. Use years 1870 to 2019 (need full years of data; no missing). Select plot wavelet
Create an autocorrelation plot?  
Select the NOAA/PSL PDO (pregenerated time-series). Set time to 1892-2019. Select cross/auto correlation as plot type.

Referencing WRIT Products

You can use the plots produced from this page in publications. We ask that you acknowledge the NOAA Physical Sciences Laboratory in the publication. For example,
"Image provided by the NOAA Physical Sciences Laboratory, Boulder Colorado from their web site at https://psl.noaa.gov/".
You will also need to reference the datasets directly. For references/citations please go to the Reanalyses.org ocean or atmospheric wiki pages for a current list of reanalysis and citations. For other datasets and time-series, there are links from the WRIT output pages and dataset documentation webpages.