Fisheries Research 101 (2010) 108–115 Contents lists available at ScienceDirect Fisheries Research journal homepage: www.elsevier.com/locate/fishres Improved estimation of trawling tracks using cubic Hermite spline interpolation of position registration data Niels T. Hintzen ∗ , Gerjan J. Piet, Thomas Brunel IMARES, Wageningen UR, Institute for Marine Resources and Ecosystem Studies, PO Box 68, 1970 AB IJmuiden, The Netherlands article info Article history: Received 23 February 2009 Received in revised form 24 September 2009 Accepted 28 September 2009 Keywords: VMS Interpolation Fishing impact High resolution Hermite spline abstract For control and enforcement purposes, all fishing vessels operating in European waters are equipped with satellite-based Vessel Monitoring by Satellite systems (VMS) recording their position at regular time intervals. VMS data are increasingly used by scientists to study spatial and temporal patterns of fishing activity and thus fishing impact (e.g. surface of sea bed trawled during a fishing trip). However, due to their low resolution (2 h basis), these data may provide a biased perception of fishing impact. We present here a method aiming at interpolating vessel trajectories from VMS data points to obtain higher-resolution data on vessel trajectories which in turn should provide improved estimates of the spatial and temporal patterns of fishing activity and hence fishing impact. This method is based on a spline interpolation technique, the cubic Hermite spline (cHs), using position, heading and speed to interpolate the trawl track of a vessel between two succeeding VMS data points. To take uncertainty of the interpolated track into account, the method also determines a confidence interval, which represents the spatial distribution of vessel presence probability between two successive VMS positions. The cHs method was compared to the straight line interpolation technique using a reference data set with intervals of 6 min which was assumed to represent the real trawl tracks. The results showed that the cHs method approximates the real trawl track markedly better than a straight line interpolation. The cHs method should therefore be preferred to the conventional straight line approach to interpolate vessel tracks in studies aiming at estimating fishing impact from VMS data. © 2009 Elsevier B.V. All rights reserved. 1. Introduction Originally, the Vessel Monitoring by Satellite system (VMS) was introduced by the European Union in 2002 (EC, 2002) for control purposes to check, among others, whether fishing vessels are oper- ating in areas where fishing activity is not allowed or to check the fishing activity of vessels that hold quotas and licences to fish in specific areas. VMS data is a valuable source of information for fish- eries scientists and have manifold potential applications: analyzing the dynamics of fisheries (Kourti et al., 2005), providing high reso- lution information on fishing effort (Mills et al., 2007) or describing fish distribution (Bertrand et al., 2005). However, most often VMS data is only provided at intervals of one or two hours (Bertrand et al., 2005; Kourti et al., 2005; Mills et al., 2007). This may not be precise enough for some applications, e.g. estimating vessel trawl track (Deng, 2005) or study the impact of fishing gear on the ben- thic fauna (Piet et al., 2000; Hiddink et al., 2006a,b). Information on a higher resolution is especially needed to accurately account for the effects of trawling on more sedentary ecosystem components ∗ Corresponding author. Tel.: +31 317487090; fax: +31 317487326. E-mail address: niels.hintzen@wur.nl (N.T. Hintzen). such as benthic organisms (Piet and Quirijns, 2009) and to measure fleet responses to management actions (Mills et al., 2007). A possi- ble solution for this problem is to interpolate between the position registrations, resulting in improved high resolution estimates of spatial fishing patterns. Most research on interpolating trajectories is carried out in the field of animal tracking experiments (Jonsen, 2003; Ryan, 2004; Jonsen, 2005; Tremblay, 2006; Hedger et al., 2008) where several different types of techniques, such as state-space modeling, ran- dom walk approaches and spline interpolations, have been used to either capture animal behavior or to reconstruct their move- ment patterns. Most of these studies use GPS positioning data as their main source of information. However, in fisheries, only few have attempted to capture fishing vessel behavior based on tracking data like VMS. In these attempts fishing impact was mostly repre- sented by the VMS data points themselves (Rijnsdorp et al., 1998; Dinmore, 2003; Hiddink et al., 2006a,b; Piet et al., 2007). But some recent studies have focused on interpolating trawl tracks from GPS positioning data. In this area of research, linear interpolation (con- necting succeeding data points with a straight line) is commonly applied and only the width of the gear is accounted for in recon- structing a trawled surface (Eastwood, 2007; Stelzenmuller et al., 2008). However, Deng (2005) stated that straight line interpolation 0165-7836/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.fishres.2009.09.014