A model for the translation of South African economic activity into shipping container demand

Van Eeden, Joubert (2018-03)

Thesis (PhD)--Stellenbosch University, 2018.

Thesis

ENGLISH ABSTRACT: Quay wall container forecasts are often done by broad-stroke methods with large-scale infrastructure decisions then based on these forecasts. This research problem requires an investigation into more accurate long-term forecasting methods. A mixed methods research design was followed, combining quantitative and qualitative data, with the primary objective to establish design requirements for and to develop a content-based quay wall container framework. Secondary objectives were to also establish design requirements for quay wall modelling frameworks for transhipment and empty containers. These secondary objectives originated from exposure to literature and data from the primary research. The aim of this dissertation was to redefine the importance and usefulness of content-based container forecasting techniques. This would enable port planners to base their container volume forecasts on economic activity, i.e. validated demand, and not on the perceived reality of historic containers or other broad indicators. The mixed method design combined literature on container modelling techniques, demand-side and supply-side container shipping factors and the impact of port networks on quay wall container volumes. The literature showed only a few scholars venturing into the field of high granularity container forecasting methods. Those that did propose methods used mostly derivatives of traded commodities like GDP, trade, or population growth as input drivers. Many scholars referred to, but very few used container contents in their modelling, mostly due to data unavailability. Rich secondary datasets received from various parties, i.e. TNPA, SARS, TFR and shipping lines, were all instrumental in understanding the relevant parameters. All datasets contributed in their own way to the development of the final set of parameters. To support this secondary data, primary research was conducted with freight owners, industry associations, LSP’s, shipping companies, port authorities and terminal operators via a survey and focus groups. Feedback from survey respondents and focus groups confirmed the user requirements identified earlier. It also confirmed the importance of the identified requirements and the inputs that were obtained from analysing container content data. Design requirements were consolidated from all the mixed methods research inputs. Key parameters to forecast full container volumes across the quay wall are: • Spatial disaggregation to define outputs per international geographic region and per port; • Rate of containerisation of each commodity; • Commodity port preference; • Physical container types; • Weight of commodity per physical container type. The container modelling frameworks and modelling process for three functional typologies, were developed based on the design requirements. The inputs, parameters, modelling process, forecasting influencers and outputs for each of the defined functional typologies were discussed separately with a confidence level for each of the aspects. The confidence levels provides an indication of the current status of the parameter values and provides guidance towards future improvement areas. The container modelling frameworks went through a verification and validation process. The proposed model is expected to provide a more accurate container forecast to port infrastructure planners. Using these drivers in forecasting models will inform port planners with validated demand towards calculated decisions on initiating port container infrastructure projects at the right moment in time.

AFRIKAANSE OPSOMMING: Kaaimuur houer vooruitskattings word dikwels gedoen deur breëstrook metodes met grootskaalse infrastruktuurbesluite wat dan gebaseer is op hierdie voorspellings. Hierdie navorsingsprobleem vereis 'n ondersoek na meer akkurate langtermynvoorspellingsmetodes. 'n Navorsingsontwerp met gemengde metodes is gevolg, deur kwantitatiewe en kwalitatiewe data te kombineer, met die primêre doelwit om ontwerpvereistes vir en die ontwikkeling van 'n inhoudsgebaseerde kaaimuurhouerraamwerk te vestig. Sekondêre doelwitte was om ook ontwerpvereistes vir kaaimuurmodelleringsraamwerke vir oorladings- en leë houers te vestig. Hierdie sekondêre doelwitte het natuurlik ontstaan uit blootstelling aan literatuur en data uit die primêre navorsing. Die doel van hierdie proefskrif was om die belangrikheid en bruikbaarheid van inhoudsgebaseerde voorspellingsmetodes vir inhoud te herdefinieer. Dit sal hawebeplanners in staat stel om hul houervolume voorspellings op ekonomiese aktiwiteit te baseer, dit wil sê gevalideerde vraag, en nie op die waargenome werklikheid van historiese houers of ander breë aanwysers nie. Die gemengde metode ontwerp kombineer literatuur oor houermodelleringstegnieke, vraagkant- en aanbodkant-houer verskepingsfaktore en die impak van hawe netwerke op kaaimuurhouervolumes. Die literatuur het slegs 'n paar aktiewe navorsers getoon in die veld van hoë diepte houer voorspellingsmetodes. Diegene wat metodes voorgestel het, het hoofsaaklik afgeleides van verhandelde kommoditeite gebruik, soos BBP-, handels- of bevolkingsgroei as insetdrywers. Baie navorsers verwys na, maar baie min gebruikte houerinhoud in hul modellering, meestal weens die onbeskikbaarheid van data. Ryk sekondêre datastelle wat van verskillende partye ontvang is, soos TNPA, SARS, TFR en skeepsrederye, was instrumenteel in die verstaan van die relevante parameters. Alle datastelle het op hul eie manier bygedra tot die ontwikkeling van die finale stel parameters. Om hierdie sekondêre data te ondersteun, is primêre navorsing gedoen met vrageienaars, bedryfsverenigings, Logistieke diensverskaffers, rederye, hawe owerhede en terminaal operateurs via 'n opname en fokusgroepe. Terugvoer van opname respondente en fokusgroepe het die gebruikersvereistes wat voor hierdie gebeure geïdentifiseer is, bevestig. Dit het ook die belangrikheid van die geïdentifiseerde vereistes en die insette wat verkry is om die inhoud van die houerinhoud te ontleed, bevestig. Ontwerpvereistes is gekonsolideer uit al die navorsingsinsette vir gemengde metodes. Sleutelparameters om volhouervolumes oor die kaaimuur te voorspel, is: • Ruimtelike disaggregasie om uitsette per internasionale geografiese streek en per hawe te definieer; • Verhouding van houerverpakking van elke kommoditeit; • Kommoditeits hawe voorkeur; • Fisiese houertipes; • Gewig van kommoditeit per fisiese houertipe. Die raamwerkmodelleringsraamwerke en modelleringsproses vir drie funksionele tipes, is ontwikkel op grond van die ontwerpvereistes. Die insette, parameters, modelleringsproses, voorspellings-invloede en uitsette vir elk van die gedefinieerde funksionele tipes is afsonderlik bespreek met 'n vertroue vlak vir elk van die aspekte. Die vertroue vlakke gee 'n aanduiding van die huidige status van die parameterwaardes en bied leiding oor toekomstige verbeterings. Die houermodelleringsraamwerk het 'n verifikasie- en valideringsproses gevolg. Die model sal na verwagting 'n meer akkurate houervoorspelling vir haweinfrastruktuurbeplanners voorsien. Die gebruik van hierdie drywers in voorspellingsmodelle sal hawebeplanners van gevalideerde vraag inlig om berekende besluite te neem oor die inisiëring van hawe houerinfrastruktuurprojekte op die regte oomblik in tyd.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/103296
This item appears in the following collections: