Return to Climate ChangeClimate Data Quality

How important is climate data quality?

  • For some climate analysis, e.g. climate change detection, additional quality control is required.
  • Variations in instrument type over the years have introduced inconsistencies in climate data records resulting in non-climatic 'jumps' and artificial trends. Other changes such as construction of nearby buildings or growth of vegetation cause similar affects.
  • These artificial changes are often of a size comparable to natural climatic variations, and so they make the detection of real climatic trends very difficult. A change in instrument shelter around the turn of the century at many Australian sites was associated with a sudden 'jump' in recorded temperatures. An example of an inadvertent "change" in climate is shown below. When the Lord Howe Island site relocated about 500 metres closer to the coast in the early 1950s, recorded daytime maximum temperatures suddenly fell and overnight minimum temperatures increased causing artificial trends in the historical records.

Lord Howe Island - maximum temperatures

  • Procedures to identify and adjust for these non-climatic influences generally involve
    • analysis of historical station information,
    • visual examination of climate records for potential problems, and
    • application of statistical tests to provide an estimate of artificial biases and confidence in their detection.
  • Steps are in place to ensure that consistent climate observations are maintained with the introduction of new instrumentation. This is particularly important given the increase in the number of Automatic Weather Stations (AWSs) in the observation network.
  • A Climate Data Quality Issues (CDQI) group has been established to identify and address issues which have the potential to cause deleterious impacts on the homogeneity and continuity of the Australian climate record. These aims will be achieved through
    • contributing to the maintenance of the Reference Climate Station (RCS) network;
    • monitoring observations, coding practices, and data archiving policy so that the climate data archive is suitable for climate monitoring and research;
    • conducting studies into climate data quality including homogeneity issues (which are primarily caused by the effects of changes in site and instrumentation) particularly for the Reference Climate Station (RCS) network; and
    • recommending comparison experiments between different instrumentation to provide data for homogeneity assessment.