The Movement interface allows users to enter parameters that determine how fast mobile communities can move.  There are also several options that inform how and where movement occurs.  Movement parameters are defined per community-class or for all (“..All..”) otherwise undefined classes.  Movement can also be related to a variable and season.  Interpolation options include linear and step (Figure).  For example, “human-adults” and “human-kids” could share or have different sets of maximum velocities that decrease with elevation and are lowest in the winter because those humans wear more clothes in winter.

Figure.  Movement rates are related to a variable and are seasonal.


Mobile communities move to areas of greater attraction.  The concept of attractions is described further in the logic section, but, in short, attractions are generated by things like forage and instinct and attractions are added together to form a total attraction layer that guides motion. 

The Movement interface includes several options some of which affect attractions and some which affect motion.  The options that affect attraction can be used to limit the spread of attractions to occur within selected regions (e.g., wildlife corridors) or mediums (e.g., water).  Additionally, attractions can be spread or adjusted according to suitabilities, which are essentially a paired data set that relates a variable to a suitability for a particular community or community-class.

The options that affect motion are a choice of whether the community should move whenever a more advantageous (higher attraction) location is available or whenever a location is available that is a user-specified % or more advantageous than the current location.  Movement can also occur using relative velocities.  When that option is on motion occurs relative to the magnitude and direction of a user-selected vector data type or set.  There is also a dropdown list with four options for movement, maximum of potential gradients, weighted average (density given priority), linear combination using rule weights, and maximum enforcing density rules.  Each option uses a different approach when determining gradients from already computed attractions (mobile communities move per the steepest local gradients – the direction of the highest immediate increase in attraction).

The Linear combination using rule weights option computes gradients directly from the total attraction layer described above.  This is the most straightforward option.  It computes gradients based on defined rules with no other considerations.  It is recommended as the default option.  The other options introduce different rationale. 

The Maximum of potential gradients option computes gradients based on attractions aggregated to the rule level and then picks the maximum of those rule-aggregation attractions as the local gradient to guide movement.  In other words, if two rule types are generating attractions, perhaps instinctual and forage, movement will be based on the higher gradient of the two rule types.  So, if instinctual rule attractions were generating a gradient that would encourage motion to the northeast and forage rule attractions were generating a steeper gradient to the northwest (and total attractions were generating a gradient to the north), the community would move to the northwest.  The consequence of this approach is that communities may move up the highest gradient and stay at an attraction peak, ignoring the attractions of other rules, until the attraction that created the highest gradient weakens.

The Weighted average (density given priority) option was developed to favor consideration of density attractions.  The final gradient to guide motion was set equal to the gradient of the density rule-aggregation attractions plus the product of one minus the absolute value of the density gradient times the gradient of all other rule-aggregation attractions ((1 - |density gradient|) * other rules gradient)).  Numerically, this means that as the mobile community neared ideal density conditions, the density gradient would decrease, and the collective gradient of other rules would become increasingly important.

The Maximum enforcing density option uses the same numeric approach as the Maximum of potential gradients method where attractions are aggregated to the rule level, gradients are based on those aggregations, and motion is based on the highest rule-based gradient.  The difference for this option is that gradients based on the density rule are used whether that rule has the maximum gradient or not.

It is important to recognize that there is no single correct mode of motion for ecological communities.  Scientific literature includes experiments that demonstrate organisms’ ability to geolocate and learn.  EFMSim is constructed to allow other ideas for motion to be coded and applied in simulations.