SA Modclllllg the Spatial and Temporal DIstribution of Lanice cOl1chi/eg{l
IS I
ent areas with different densities of L. col/chilega. These can help to validate the
model assumption that the local density of L. conchilega is strongly influenced by
the overall velocity of the near-bottom flow: low density occurs in areas where low
velocities prevail. whereas high densities occur where high velocities prevail.
Measurements of flow velocity, which were performed in June 1994 on the
Groninger Plate, indicated a mean velocity of about 10 cm/s in an area with a low
density stand of L. conchilega, and 20 cm/s near high density patches (Brandt et al.
1995). These results tentatively confirm the model assumption. For a real test,
however. much more detailed records of the flow regime are required. It must be
emphasized that in our phenomenological approach "low velocity" is a metaphor
for one or more aspects of the near-bottom tlow which are not yet known (cf.
Friedrichs 1996).
We do not claim that the model provides a detailed retlection of reality. as it
merely takes into account those factors which are thought to be essential for the
spatial and temporal distribution of L. conchilega. The model assumptions are
extremely simple because more detailed information is unavailable (but see
Friedrichs 1996) and because unnecessary detail would make it more difficult to
explore and understand what the model does. However, this all means that the
model cannot be tested directly. Instead. it allows testable hypotheses to be formulated. The model gave us an idea of what might be essential in reality, and suggests the following hypothesis:
The regional distribution pattern of L. conchilega is not caused by local differences in larval supply. This hypothesis goes back to the linear model (Heuers et al.
1998). which could not reproduce the spatial and temporal dynamics recorded by
Hertweck (1995) by solely varying larval supply (maxL). The records of larval
abundance and phenology taken in the ELA W AT project give no indications that
larval supply on different parts of the Groninger Plate were different. To further
test this hypothesis. driftnets should be located in areas with low or high tube density respectively.
The model can also be used to formulate a verbal model which explains the
striking spatial pattern of mounds and depressions emerging in some parts of the
sandtlats. The mounds are densely populated with tubes of L. conchilega, whereas
the depressions. which are filled with water even at low tide, are free of tubes. The
typical span of the mounds and depressions is one to three metres. i.e. larger than
the mounds-and-depressions pattern which also emerges in some parts of the sandflats. but without being populated by L. conchilega (Heuers et al. 1998). The verbal model states that dense aggregations of tubes detlect the near-bottom tlow.
This may lead to such high lateral tlow velocities that a further lateral expansion of
the patches as predicted by the model is prevented. Similar phenomena have been
reported for macroalgae in rivers (Sand-Jensen & Mebus 1996).
Although models cannot supplant field studies and experiments, they help to
integrate the knowledge available and to identify important gaps in this knowledge.
The results presented here merely represent some initial, preliminary runs of the
model. Nevertheless. we believe that models of the kind presented here are an
ideal tool to assist the exploration of spatial and temporal dynamics of
L. cOf/chilega and other benthic species.
IS I
ent areas with different densities of L. col/chilega. These can help to validate the
model assumption that the local density of L. conchilega is strongly influenced by
the overall velocity of the near-bottom flow: low density occurs in areas where low
velocities prevail. whereas high densities occur where high velocities prevail.
Measurements of flow velocity, which were performed in June 1994 on the
Groninger Plate, indicated a mean velocity of about 10 cm/s in an area with a low
density stand of L. conchilega, and 20 cm/s near high density patches (Brandt et al.
1995). These results tentatively confirm the model assumption. For a real test,
however. much more detailed records of the flow regime are required. It must be
emphasized that in our phenomenological approach "low velocity" is a metaphor
for one or more aspects of the near-bottom tlow which are not yet known (cf.
Friedrichs 1996).
We do not claim that the model provides a detailed retlection of reality. as it
merely takes into account those factors which are thought to be essential for the
spatial and temporal distribution of L. conchilega. The model assumptions are
extremely simple because more detailed information is unavailable (but see
Friedrichs 1996) and because unnecessary detail would make it more difficult to
explore and understand what the model does. However, this all means that the
model cannot be tested directly. Instead. it allows testable hypotheses to be formulated. The model gave us an idea of what might be essential in reality, and suggests the following hypothesis:
The regional distribution pattern of L. conchilega is not caused by local differences in larval supply. This hypothesis goes back to the linear model (Heuers et al.
1998). which could not reproduce the spatial and temporal dynamics recorded by
Hertweck (1995) by solely varying larval supply (maxL). The records of larval
abundance and phenology taken in the ELA W AT project give no indications that
larval supply on different parts of the Groninger Plate were different. To further
test this hypothesis. driftnets should be located in areas with low or high tube density respectively.
The model can also be used to formulate a verbal model which explains the
striking spatial pattern of mounds and depressions emerging in some parts of the
sandtlats. The mounds are densely populated with tubes of L. conchilega, whereas
the depressions. which are filled with water even at low tide, are free of tubes. The
typical span of the mounds and depressions is one to three metres. i.e. larger than
the mounds-and-depressions pattern which also emerges in some parts of the sandflats. but without being populated by L. conchilega (Heuers et al. 1998). The verbal model states that dense aggregations of tubes detlect the near-bottom tlow.
This may lead to such high lateral tlow velocities that a further lateral expansion of
the patches as predicted by the model is prevented. Similar phenomena have been
reported for macroalgae in rivers (Sand-Jensen & Mebus 1996).
Although models cannot supplant field studies and experiments, they help to
integrate the knowledge available and to identify important gaps in this knowledge.
The results presented here merely represent some initial, preliminary runs of the
model. Nevertheless. we believe that models of the kind presented here are an
ideal tool to assist the exploration of spatial and temporal dynamics of
L. cOf/chilega and other benthic species.