A goal of the NASA STEREO mission is to study the feasibility of forecasting
the direction, arrival time and internal structure of solar coronal mass
ejections (CMEs) from a vantage point outside the Sun-Earth line. Through a
case study, we discuss the arrival time calculation of interplanetary CMEs
(ICMEs) in the ecliptic plane using data from STEREO/SECCHI at large
elongations from the Sun in combination with different geometric assumptions
about the ICME front shape (Fixed-\Phi (FP): a point and harmonic Mean (HM): a
circle). These forecasting techniques use single-spacecraft imaging data and
are based on the assumptions of constant velocity and direction. We show that
for the slow (350 km/s) ICME on 2009 February 13-18, observed at quadrature by
the two STEREO spacecraft, the results for the arrival time given by the HM
approximation are more accurate by 12 hours than those for FP in comparison to
in situ observations of solar wind plasma and magnetic field parameters by
STEREO/IMPACT/PLASTIC, and by 6 hours for the arrival time at Venus Express
(MAG). We propose that the improvement is directly related to the ICME front
shape being more accurately described by HM for an ICME with a low inclination
of its symmetry axis to the ecliptic. In this case the ICME has to be tracked
to > 30{\deg} elongation to get arrival time errors < 5 hours. A newly derived
formula for calculating arrival times with the HM method is also useful for a
triangulation technique assuming the same geometry.
