You need a more modern web browser, sorry!
               Zoom       Scroll
  Location: Rural / Semi / Urban     Night light: Rural / Semi / Urban

You need a more modern web browser, sorry!
Use these controls to limit the stations used when the *Align stations on common baseline* checkbox is selected. Selected sites that do not meet the criteria are colored purple.
  • First, set the baseline period
  • Then set the range to limit the first set
  • Use the range fields several times so you can require a few years near one date and a few near another
  • To reset the flags, set a new baseline period

    FunctionFromToWith at Least this many
    measurements within the range
    Click button to apply
    Set baseline
    Require data
    Require data

For instance, of the 5,255 available GHCN adjusted sites,

  • Of those, only 4567 have 15 years of data between 1961 and 1990
  • Of those, only 1000 have  8 years of data between 1900 and 1910
  • Of those, only   520 have  8 years of data between 2000 and 2010

To display individual sites in the lower plot, select *station* and click on the site. To display purple (restricted) sites in the lower plot, turn off *Align stations on common baseline*.


    BELGRANO This base was moved, but the 8C temperature increase was not corrected
    TULCAN/EL ROS There is a major 10°C discontinuity in the data
    Ship data
    There are 7 ships included with the land data.
    Only ship M has enough data to be of any value.

Station data

    ID set the map selection mode to Station The first 3 digits indicate the country
    Name click on a staion to populate this  
    Pop Type Rural/Semi/Urban - Station type by size of nearest town
    Light Type Rural/Semi/Urban - Station type by light seen from space
    min avg max  
    0 0 0
    ghcn plots These show raw and adjusted data
    Search Berkeley Earth There is no simple translation to their ID's - so search

Plot Controls

    Set plot limits
    Control Charty max
    y min
    x minx max

    Toggle the chart selections to apply new values
Stations plot

This is good for looking at small regions and/or countries

Averages plot


Trend line formulas - currently, only on Hi/Low graph
    Just fit a line to the end points
    y = (y0 + 273.15)exp( (t-t0)/k ) - 273.15
    y = (y0 - ymin + 0.1)exp( (t-t0)/k ) + ymin - 0.1

The Null Hypothesis is that the temperature change is indistinguishable from zero. For this to be true, ....

Load Local Data

Load data files from local storage. While the button shows all the files in a directory, only xml files with the predefined names will be read. Therefore, you can simply select them all. Once a file is loaded, it will not try and load it again.

Because of how javascript security is implemented, there is no way to load data from a local storage device unless the user manually selects the files via an input files control .. which is what this page provides.

Standard deviation (σ)

This applies only to the plot with the number of stations (sites) chart displayed at the bottom

The values are not meaningful if the average is "Corrected for station distribution"

To produce an approximately unbiased estimator of the standard deviation of an average of averages, the divisor should be (N - 1.5)
    N is the number of stations selected
    Creates a biased estimate
    Creates an unbiased estimate when applied to a variance or a mean
    Approximates an unbiased estimate when applied to a standard deviation of averages


    Select an alternate map
    This is the default map
    Deserts, Forests, Ice, and more ref
    The Earth at night ref

UCAR and Keeling CO2 Data

The UCAR Simple Climate Model predicts how much the temperature will increase for an increase in CO2 - however, their data actually demonstrates that the model is wrong. By default, only the data after 1900 is shown in the Hi/Lo graph and, with a sensitivity of 3 or 4 and using 1975 as the base year, the agreement between the theory and the actual data is pretty good. However, the complete UCAR "historical data" is a poor match to the theory - when data from the previous century is included (use the checkbox), the correlation is terrible.

The Keeling data (1958 to present - only the August averages are plotted) is almost identical to the UCAR data for that period.

Sensitivity - ΔT when CO2 is doubled