Simulating Roman Trade Patterns

Figure one ‘Course taken when tracking in to the Wind’

I am a professional software developer and also studying for a PhD part time at the University of Southampton. My main interest at the moment is in the simulation of different routes connecting Portus to the other ports of the Mediterranean – topics discussed way back in Week One of the course e.g. in the Links to Other Ports step. My work uses a combination of computational approaches. I produce maps and display them in a Geographic Information System (PostGis, QGIS and JTS Topology Suite) . These are open source tools for managing spatial data such as modern wind and tide patterns, ancient coastlines (derived in part from the kinds of bathymetric data capture referred to in Week two), locations of ports (linking to the work discussed by Leif in Pelagios: a Sea of Connections).

I also use computational methods designed to take these data and turn them into simulations of where ships might sail under given conditions (including what we know about the capabilities of ships from the work referred to by Julian post on Roman Mediterranean Shipping). Finally having calculated potential routes I simulate the ‘cost’ in terms of time that a ship might use in undertaking a particular voyage. This work is drawing on modern data and also methods use in modern navigation, and also on the diverse archaeological evidence we have from Portus and elsewhere that provides the framework of known connections. For example, the pattern of amphorae discussed by Simon and also in Pina’s PhD as discussed in her post on Ceramics as material culture: Study of the North African amphorae from Portus.

Roman merchant ships used wind power as their source of power, by simulating the possible movement patterns of a ship its possible to create a theoretical model of how goods could be distributed from Portus and other ports throughout the Empire.

With the wind blowing in the right direction the maximum speed can be achieved but to sail against the wind, a ship has to ‘tack in the wind’ by steering a zig zag course, see figure one above.

When a ship was sailing with the wind (a favourable wind) the distance covered is X (Red line figure 2) then the distance sailed against the wind( a foul wind) Y (Blue line figure two)is the horizontal length of a sine wave of length X. The ratio of the best and worse distances (X&Y) may be used to deal with the conditions when both foul and favourable winds are encountered, assuming that 90 degrees is the best wind and 270 degrees is the worst wind, then 0 and 180 degrees would be a ratio of 50% of the best speed and 50% of the worse wind.

Figure two 'Ratio of worst speed against best speed' — Figure two ‘Ratio of worst speed against best speed’

Theoretically this creates a carotid shape, where the border represents what a ship could be, see Figure three, where:

The red point represents the starting position of the ship.
The blue area represents the area that could be covered from the starting point if the best possible was available in all directions.
The yellow area presents the area that could be covered from the starting point if the worse wind blew constantly.
The green area presents the ratio of the blue and yellow areas and represents the maximum distance that could be covered assuming the wind was blowing at 45 degrees.

Figure three 'Unit movement carotid' — Figure three ‘Unit movement carotid’

The prevailing winds in the Mediterranean basin in the summer blow towards Africa, see figure four.

Figure four 'Prevailing winds' — Figure four ‘Prevailing winds’

By tessellating the unit movement (figure three), using the prevailing wind (figure four) to orientate it creates an isotropic curve that show the possible distances that could be covered by a ship, using the best possible course, see figure five.

Figure five 'Tessellated movement carotid' — Figure five ‘Tessellated movement carotid’

By extending this method to the entire Mediterranean basin, it possible to map how long it would take to distribute goods from Portus to other areas of the Empire, see figure six. Where Blue dots denote a known Roman Port (a port can be anything from a good beach to Portus) and Green lines represent different generations in 8 hour periods

Figure 'Six Isotropic curves from Portus' — Figure ‘Six Isotropic curves from Portus’

To sail a ship efficiently requires constant adjustment of the sails, correct operation of the steering oar and steady winds. By assuming that a crew could capture a fix percentage of the available wind and applying to this the size of the unit movement, it’s possible to simulate the results of different movement patterns, see figure seven.

Figure seven 'Numbers of different ports reached against periods sailed' — Figure seven ‘Numbers of different ports reached against periods sailed’

By using known sources of goods in the Roman and applying the above model, it’s possible to create a supply map for resources considering the supply points for Olive Oil, see Figure eight, where the brown area are the know supply areas for olive oil.

Figure eight 'Supply map for Olive Oil Resources' — Figure eight ‘Supply map for Olive Oil Resources’

As would be expected oils from Tunisia and Algeria should be expected to be present at Portus because they are the nearest resource. But it is interesting to note that supplies from North Africa would supply equal supply Crete as supplies from ‘Syria’ area.