ZONTEM (short for Zonal Temperatures) is a method for computing the change in global temperature over the recent period (1880 CE to present) from published records of monthly mean temperatures (GHCN-M V3).
ZONTEM aims to be as simple as possible while still giving a reasonable result.
You will need to be able to run Python from the command line. We require Python version 2.
This downloads the GHCN-M dataset (of monthly average temperatures) as a compressed file, unpacks it, and runs the ZONTEM analysis.
The result, a CSV file, is placed in the
output/ directory. It
has a name based on the name of the input file used, which
changes each month. Mine is called
Running ZONTEM again
run-zontem.py will not download the compressed GHCN-M file
again if it sees it in the
input/ directory. So if the file is
there but broken (maybe due to an interrupted download), or you
want a more recent version of GHCN-M (a new version is published
roughly every month), then you should empty the
The actual analysis is done by the program
you can run that directly:
The input is a number of records, each record being a time series of monthly (air) temperature averages from a single station.
The input records are distributed into N (=20 by default) zones according to the latitude of the station. Each zone represents the surface of the globe between two circles of latitude; each zone covers an equal area.
All the station records in a zone are combined into one record by the Reference Station Method described in [HANSEN1987].
All the zone records are combined into a single global record using the same method.
The global record is converted to annual anomalies by first converting to monthly anomalies and then averaging into years where a year has all 12 monthly anomalies present.
The most obvious competing analyses of global temperature change are GISTEMP, CRUTEM4, NOAA (this is not an exhaustive list).
data sources. ZONTEM uses the GHCN-M product as its only data source. Other analyses use this supplemented with SCAR READER records, various priv comm records, or replace GHCN-M with a privately maintained database with similar coverage.
SSTs. Other analyses may have an optional procedure where Sea Surface Temperatures (SST) are used over the ocean. CRUTEM4 has a sister product, HadCRUT, that also incorporates SSTs. Whilst recognising that the Earth's surface is mostly ocean, ZONTEM does not incorporate SSTs.
gridding. Other analyses compute global temperature anomaly via a gridded product (the gridded product may be regarded as more essential than the global summary). ZONTEM dispenses with the grid, instead using a small number of zones (you could alternatively think of it as a 1x20 grid).
QC. Other analyses may have a QC step that rejects invalid station records. ZONTEM assumes that the input has been quality controlled.
inhomogeneity. Other analyses may have a step that adjusts or rejects inhomogeneous records. ZONTEM does not perform such a step (but usually uses the adjusted version of GHCN-M as input).