Program description HISKLIM

  • Introduction
  • Data-infrastructure and data-archiving
  • Digitization
  • Quality control and homogenization
  • Schematic overview of HISKLIM (PDF-file, 25 kB)

    •  
     

    Introduction

    Long high-quality historical time series and databases are mainly needed to achieve clarity about anthropogenic climate change in relations to natural climate variability. Worldwide there is an increasing request for such time series and databases with a high temporal resolution. Besides, there is a growing awareness of the importance of distributing these data publicly by a well accessible medium like Internet. The two most important objectives of HISKLIM are strongly connected with these: (1) improving the quality of existing long historical climate time series and databases; and (2) making historical climate data publicly available in a user-friendly manner (data-infrastructure). Furthermore, data-rescue and data-archiving play an important role.
     

    Data-infrastructure and data-archiving

    The data-infrastructure for historical data, both maritime and land, doesn’t satisfy the requirements of these days. This applies to other types of data (satellite, radar, model, etc.)  as well. Therefore, at KNMI a process started with the objective to improve the data-infrastructure. An important problem with historical climate data is the insufficient accessibility of the accompanying metadata. Furthermore, there is no written KNMI archiving policy with respect to digitized and non-digitized observations and there is no clear policy with respect to the distribution of historical climate data and the protection of the data (e.g. backups).

    What do we want to achieve?
    HISKLIM strives for a well accessible (by internet) and mastered metadata information system of all historical data (whether it is digitally available or not). For the user the system works as a search-catalogue. It forms the basis for on-line downloading of various (to be determined afterwards) historical observations. The system will be set up such that it can become a National Climate Database (NKD) containing also non-HISKLIM data. Although our final goal is approaching observations in databases, at first downloading of selected observations may suffice. In addition a data policy will be formulated stating which type of data and with what kind of restrictions may be downloaded. Reckoning with the KNMI-Catalogue, the system will be specified according to ‘free flow of data’, inclusive a user-friendly manner to approach the system. For the pre-1850 maritime data we will investigate the possibility to setup an international Maritime Climate Database.

    There is a clear need for a (written) archiving policy for digitized observations (storage of observations with high temporal resolution, backup strategy, etc.) and non-digitized observations. Because of this HISKLIM wants to setup an archiving policy such that: (1) it is clear for everyone which observations are stored and which not; (2) observations are easy to trace; (3) observations are stored safely and effectively; and (4) it is clear for everyone which persons are responsible.
     

    Digitization

    It appeared from an inventory of maritime data and landdata that a huge amount of observations is still slumbering in the archives. To make these data available for climate research, etc., the first step is to digitize the observations. By digitization we mean something different then libraries and archives usually do. Libraries and archives denote filming and digitally photographing of sources as digitizing. In that case the data will not be available in ASCII. For us the data need to be in ASCII in order to be able to make calculations with the data. The latter way of digitizing is, however, a labor-intensive job that requires discipline and endurance to a large extent. Fortunately, for the greater part this is a onetime event, provided that the original observations are stored without corrections beforehand.

    What do we want to achieve?
    Our objective is to digitize as much as possible of the maritime data and land data that slumber in the archives.  For that we want to deploy a sufficient number of persons, both internally and externally, such that the progress of this work is guaranteed. Not every data source has same high priority to be digitized within HISKLIM. On the other hand, these sources may contain climate data that hardly available nowadays (e.g. 1—minute rainfall time series from strips) and may be needed in the future. To digitize the data within a reasonable term, much depends on the possibilities to obtain subsidies.
     

    Quality control and homogenization

    Quality control and homogenization are both activities that aim at the improvement of the quality of climate time series. Quality control is an everyday routine activity at KNMI in which the observations are tested for their quality and, if needed, corrected. In this process every observation gets a quality code. For the older observations, e.g. the yearbooks of the 19th century, these codes are missing. Also the times of observation are not standardized in these books (also a topic for the point of concern for the term-stations of the 20th century). Further back into the past, the units of measurement are not standardized. Thus, for the older time series a number of operations must be carried out retrospectively.

    So far no corrections are being made for changes in measuring position, measuring instuments, etc. This implies that climate time series may contain inhomogeneities. Homogeneous climate time series are, however, essential for climate change and variability research. Therefore, homogenization of time series and guaranteeing the continuity of series is an important objective within HISKLIM. Besides, there are problems with the maritime data such as the COADS problem and the problem of making available  homogenized time series of the light vessels.

    What do we want to achieve?
    In the first place we want to correct the existing observations such that they are suitable for further distribution. These corrected time series constitute an important basis for our second objective, namely the homogenization of the Zwanenburg/De Bilt time series with the highest possible temporal resolution. Thereafter, we want to homogenize other time series as well, such as the series of the other four main stations. Furthermore we want to make sure that the continuity of existing long time series is guaranteed.
     
     
     

    intro-ne HISKLIM: program description | publications | datalinks
     
     

    Theo Brandsma