2.1 tave.aml

This AML is designed to generate monthly (average, min, or max) temperature maps using a DEM, a point coverage, and some parameters and settings in the AML itself. The AML is not very much automated. It certainly could be refined and made much user-friendlier. Please feel free to do so if you it helps your project needs.

The AML basically is written to perform a simple climate mapping procedure (see Zimmermann & Kienast, 1999). First a set of climate station’s data has to be handy. The data are assumed to be pre-analyzed (maybe summarized, etc.), and error checked. Next, least-square regressions have to be performed in order to explain temperature by elevation. The resulting intercept and (adiabatic) lapse rates are used to generate the climate maps. The following steps need to be performed or considered:

The point coverage represents the location of climate stations, and it is supposed to be called “stations”. It has temperature values stored in variables (one for each map, i.e. for each month). The values, however, are not those measured at the climate station, rather they are the 0 m.a.s.l. values, as projected from to sea-level using the (adiabatic) regression lapse rates. By this, they are unlinked from elevation and allow spatial interpolation independent of the elevation the climate variables were measured at.

The adiabatic lapse rates, which were determined by linear least-square regression, are entered in the tave.aml. Further, the AML has to be adjusted to the outline of the study are and to the grid resolution of choice. The respective coordinates and cell sizes have to be adjusted in the AML.

The AML performs the spatial interpolation of the sea-level temperature in two steps. First, it uses the temperature information and the spatial organization of the climate stations to generate a medium resolution grid. Next the spatially interpolated grid is smoothed. This is to prevent overfitting of the stations. In order to speed up this smoothing process, it is important to do this at a rather coarse cell size. Otherwise, the smoothing process would be very slow. Next the medium resolution grid of the sea-level temperatures is sampled using in accordingly spaced point lattice (using the LATTICESPOT command). Thus, point coverage has to be generated (e.g. by using the lattice.for program), so that the coordinates of the points fit the center of the medium size interpolated sea-level temperature grid. By this, the values are not interpolated during re-sampling. In a next step, the temperature values in the point lattice are spatially interpolated to the final resolution. The resolution should match the cell size and spacing of the DEM, which represents the study area. In a final step, the fine-scale sea-level temperature grids, and the equally fine-scaled DEM, and the adiabatic (regression) lapse rates are used to re-project the temperatures back to the actual DEM elevation, thus generating fine resolution temperature maps.
The AML needs some adjustments for study area outline (coordinates of the LL and UR corner of the grid) and cell sizes (for the medium and the fine scale grid), as well as for the name (&path) of the DEM. By this, the AML is not very user-friendly. Rather it is optimized to run for large areas, with a distinct knowledge of the spatial organization (outline, cell-size, etc.).
 
 
 

General specifications of the AML:

Command: &r tave      (at GRID  prompt)
Required input: DEM, station point cover, lattice cover
Output units: 1/10ºC (or F)
Speed of calculations: Rather fast due to hierarchical procedure
Flexibility of the routine: High; but user needs to adjust AML manually
User interface: No interface
Known errors:  -
Programmer N.E. Zimmermann
Download: tave.aml       (use: "save link as")
Contact: niklaus.zimmermann @ wsl.ch

 

References:

Zimmermann, N.E. & Kienast, F. 1999. Predictive mapping of alpine grasslands in Switzerland: species vs. community approach. J. Veg. Sci. 10(2).
 
 

[top] [back] [home]

Last Updated: 9/08/00
By Niklaus E. Zimmermann