The principal difference between version 2.0 and 1.0 of the MCD is that the large-scale variability model now makes use of two-dimensional, multivariate Empirical Orthogonal Functions (EOFs), which describe correlations in the model variability as a function of both height and longitude (rather than solely of height as in version 1.0). The 2-D EOFs allow realistic variability to be modelled for trajectories which span a range of longitudes. As in version 1.0, EOFs are stored for a range of latitude bands, but instead of retaining 6 1-D EOFs at each horizontal location for each of 12 seasons, now 72 2-D EOFs are stored for each latitude band and their amplitude is modelled by a set of principal components tabulated once per day (669 times) throughout the model Mars year. This procedure is no more costly to the end-user in terms of either disk storage or CPU time, but gives a much improved description of the variability as a function of both space and time and a larger variance capture.
A second difference is that two parallel datasets are now included in the climate database distribution. The first has been compiled with the same set of model properties as the version 1.0 database, although the GCMs which were used to compile the database have been extensively validated as a result of the first study and improved in several respects (Forget et al., 1999). This dataset is refered to as the ``standard dust'' dataset. A second complete database has also been prepared with an increased dust single scattering albedo in the models (see Forget et al., 1999, for details) and is refered to as the ``bright dust'' dataset. The main systematic difference between these datasets is that the ``bright dust'' data tend to have temperatures which are cooler by up to 10K in the lower atmosphere (between about 5km and 30km altitude). Differences elsewhere are less marked and are often smaller than the intrinsic model variability. An example of the differences can be seen in Lewis and Collins (1999a). The ``bright dust'' properties tend to give a better fit to a few previous observations of Martian atmospheric temperatures (Lewis et al., 1999) and could be used if only a single dataset is required. The ``standard dust'' dataset gives a better correspondance with the previous version of the database. Both may be used to check warmer and cooler extremes.
There have also been numerous small improvements and error corrections to the climate database access software since version 1.0 was released, which are now all incorporated into version 2.0. These are listed in a Changes file supplied with the software.