Geoscience Reference
In-Depth Information
Figure 2.23.
Correlation coefficients for topographic roughness/rainfall amount (monthly
averages); the roughness is generated by seven DEMs, each with a different resolution
Altitude
The correlations between rainfall amount in April and altitude, which are
considered as dependent and independent variables, respectively, measure the
intensity and the direction of the vertical precipitation gradient. The first important
point is that the correlations are positive: rainfall level increases with altitude. The
second important point is that the coefficients very little change, regardless of the
DEM which is used, and thus regardless of its resolution. This second point
highlights the fact that altitude, which has a strong influence on rainfall and
precipitation in general (all of the r coefficients are greater than 0.63 and reach up to
0.85), does not cause any scale effect. The altitude of the climatological stations is
slightly modified when the resolutions of the DEM are changed, but this has no
effect as far as the correlations are concerned [JOL 07]. The same is true for the
coefficients for December (the month which produced the minimum coefficients)
and June (the month which produced the maximum coefficients). Results from
calculations show that altitude has a uniform influence on the precipitation gradient;
irrespective of the DEM resolution the correlations remained stable for all seven
resolutions, and this was true for each month of the year.
Slope and topographic roughness
Once again the correlation for these two variables is positive. The rainfall
amount is much higher for steeper slopes and when the level of topographic
roughness is also increased. The correlations of the two independent variables show
an analog variation depending on the particular DEM in question: the relationship
between slope and roughness, and rainfall amount improves as the resolution of the
DEM gets larger, up to an optimal resolution of 2,500 m (Figures 2.22 and 2.23).
This trend highlights the fact that the monthly rainfall averages are mainly induced
by small-scale topographic factors that are present in a particular environment.
The slopes and roughness levels, which were calculated on a 50 m resolution,
are related to objects whose minimum size is approximately 100 m. Contrary to this,
slopes and roughness levels, which were calculated on large-scale DEMs (2,500 and
5,000 m), identified objects that were bigger than 7 km. The second set of small-
scale objects has a stronger influence on the spatialization of rainfall than the first
set of large-scale objects. No real change is made to this pattern, irrespective of
whether the coefficients relate to April, June (maximum coefficients) or February
(minimum coefficients).
