This is a very complex issue in contaminated land management. Projecting losses due to contamination into the future implies that the entity having to pay the compensation is trying to avoid remediation, for whichever reason.
In my opinion, past losses can be estimated by the kinds of crops planted and harvested before contamination took place. Estimation of future losses can be based reliably on past losses.
It makes a great difference if you are planting cassava with roots extending to a depth of 2-2.5 meters below ground surface [mbgs], or maize or beans with rooting depths of 0.5 mbgs or so. If we are estimating future losses, we will also have to take into account the future biological degradation of contaminants, which can only be estimated by making assumptions.
If e.g. in case of a kerosene spill on an airfield or a crude oil spill in Niger Delta phase liquid is present in the soil and on the groundwater, there will be no biodegradation, and the period of the estimate can be extended safely into the range of decades or centuries if there is no remediation.
TPH should not be the only indicator: BTEX, especially benzene and toluene are relatively highly soluble in water and will be taken up by the roots and will accumulate in oily crops like groundnut, avocado and the like. Polycyclic aromatic hydrocarbons [PAH], heavy metals, phenols etc. can be an issue as well. The spectrum of Contaminants of Concern [CoC] depends largely on the contaminating industry.
A land owner certainly would insist on being compensated for past losses, and on remediation being carried out in a manner that the land can be used without any health risk again as soon as possible.
Any mathematical “projection” or “model” of loss estimation is an estimate. And although it may be a highly sophisticated model/projection, it will still exclude a lot of known unknowns, and unknown unknowns.
I have been using the RBCA (Risk Based Corrective Action) Tool Kit and CLEA (Contaminated Land Environmental Assessment)software to estimate the impact on human health, and to determine necessary cleanup levels. This type of software is accepted by authorities in the UK and the USA, but it is based on detailed analyses, and a database containing hundreds of substances. The land use is roughly divided into agricultural, housing areas, etc. This type of software has been developed and refined during the past ten years or so, but still does not compute future developments.
Projecting historic income profiles into the future is in my opinion the correct methodology to determine losses caused by land contamination. The complex scientific methodologies of CLEA and RBCA are actually not suitable as a tool to determine compensation as they are targeted towards risk estimation and development of remediation targets, and thus are aimed at estimating remediation costs, amongst others.
Whilst this may be of interest for the polluter in terms of the “polluter-pays-principle” to determine the total costs of cleanup (of which the compensation for loss of income can be considered a part), it is of no importance for the land owner.
In addition, land owners (be it in Nigeria, Germany, USA or Tuvalu) often do not understand the technical and scientific details. The bottom line of this is: keep it simple.