2000

1000

800

600

Eq 4

400

Eq 3

200

Eq 2

100

80

60

2

4

6

8

10

20

Density . Arc Length . Width (kg /m)

eq 3. Further analysis of this data set (old and new

pendix B. Snow density (i.e., snow characteristics

CIV data) using a stepwise regression procedure

other than depth) may have been too similar be-

showed that an equation including only snow

tween tests to develop a clear effect.

depth and the ratio of arc length *a *to 1/4 the

perimeter of the wheel provided a slightly better

The initial analysis consisted of examining scat-

correlation to the data. This equation

ter plots of the independent variables (snow depth,

4a

density, tire width, sinkage and velocity) and sev-

(6)

*r*π

eral different combinations of these parameters vs.

has an *r*2 of 0.68. It is interesting that neither the

several forms of the dependent variable: rear

wheel width nor snow density is included in the

equation, even though wheel width varied from

of the trailing wheel motion resistance were con-

0.156 to 0.274 m (albeit there are only a few data

points at the narrowest width). The lack of a wheel

width effect was also suggested by the analysis of

(coefficient form), ratio of rear tire motion resis-

snow deformation by a wheel presented in Ap-

w

Undisturbed

r

a

z

Snow

h

Compacted Snow

7

Integrated Publishing, Inc. |