Viewing page 142 of 218

This transcription has been completed. Contact us with corrections.

nomenon or process involving the parameter. For example, adequate sampling of a minor constituent in the atmosphere whose lifetime is known to be years in length and which is uniformly mixed can be achieved by a very small number of sampling stations making monthly averaged measurements. If the constituent varies seasonally, the sampling frequency might be increased to weekly. On the other hand, if cloudiness is to be determined, measurements must be made globally with high spatial resolution and perhaps many times a day because of the transient nature and large variability of cloudiness.

For long-term, ongoing studies, low resolution observations might be quite adequate. However, special intensive, high resolution, high frequency measurements might be required from time to time to understand or model certain aspects of the behavior of a parameter. This intensive observation may be necessary to satisfy special requirements.

The sampling requirements for most significant geophysical parameters have been established over the years by individuals and teams studying the particular parameter/phenomenon/process in question. However, it would be very useful if these requirements were standardized in connection with the ISY, especially in the case of space-derived remote sensing data.

• Data Format

The crucial element in global standards is the development of common data formats applicable to remote sensing data as well as all additionally collected in situ and ancillary data. This request is based on the need to:

• exchange data between international teams of investigators;

• integrate remote sensing data in models of the ecosphere, the hydrological cycle, the global carbon cycle, etc.;

• integrate data collected at different scales in time and space; and

• be able to reinvestigate data collected at earlier times.

A common standard format(s) is essential for use in archiving space-derived and ancillary data sets. This will facilitate the free exchange and utilization of all geophysical data which is needed to study global change.


• Algorithm Documentation

Transforming raw satellite observations into analyzed fields of geophysical parameters is quite a complex procedure rarely satisfactorily described by instrument designers and/or agencies. From raw observations to analyzed fields, three major algorithmic steps may be identified:

• Conversion of raw data (for example, counting voltages, etc.) and housekeeping information into properly calibrated and Earth-located observations. This step includes laboratory-derived information on the instrument;

• Conversion of calibrated/earth-located data into non-analyzed geophysical parameters through a "retrieval" procedure; and

• Re-imaging the data into a regular grid.

Having obvious impact on the quality of the final analyzed fields, each of these three steps must be completely documented. To that purpose, we recommend that all necessary information be reported in a "User's Guide" divided into three parts, each corresponding to one of the items above. The last major role of such a User's Guide should be to present an accurate and complete description of the instrument. It should be immediately updated after any modification of the procedure.


• In Situ and Other Ancillary Data

Remote sensing data sets need to be complemented by in situ measurements characterizing the investigated targets, their status and distribution by biophysical and biochemical data, as well as data describing the environmental conditions. The type of in situ parameters to be collected for the "physical calibration" depends strongly on the role of the anticipated goals. It may reach from parameters describing the canopy of a single agricultural field and its underlying soil to the classification of a whole area as an agricultural use area; it may characterize single forest stands by age classes, new diversity and types of trees, to a whole region as typical forest area. These characterization parameters are collected statistically such that whole test regions can be defined. The sets of in situ data are to be complemented by additional information such as digital terrain models, climatological charts, and soil and land use maps.

Furthermore, these data sets should be complemented by different types of models to:

• Relate remote sensing data with in situ parameters;

• Relate remote sensing data with the actual topographic and environmental target location; and

• Describe the growth stage and status of targets, for instance, by successional models.


• Raw Data Archive

History has shown the immense value of preserving data sets for long periods of time. This approach enables the review of previous data sets in the light of new information or unforeseen but emerging needs. Such a review requires that we archive the raw data and the calibration and housekeeping information which was employed in the production of geophysical and biological variables. We cite here the recent case of atmospheric ozone where we have recently discovered major short-term global change. We have no

19