Fetching recent data

[universmile]

At the moment, the tables are obtained from the book Dynamics of Growth in a Finite World (1974). We should feed the model with more recent data.

[universmile]

In order to update the model with modern data, we first need to understand within World3 the following targets: (i) which types of data are used, (ii) from where the data was obtained.

We consider a first case study by looking at the Pollution system (Chapter 6).

Regarding target (i), we distinguish between Parameters, Initialisations and Time series (see the corresponding documentation https://natema.github.io/WorldDynamics.jl/stable/world3/).

Type	Description
Parameters	PYEAR, PPGF1, PPGF21, FRPM, IMEF, IMTI, FIPM, AMTI, PPTD1, PPTD2, PPOL70, AHL70, SWAT, TDD, PCTI, PD
Initialisations	T0, PPOL, PCTI
Table values	AHLM, PCRUM, POP, AIPH, AL, PCTCM, LMP1, LMP2, LFDR1, LFDR2

Regarding target (ii), the description of data is included within section 6.5 that describes the equations.

We provide hereafter an overview of which data needs an update and the requirements involved to perform properly such task.

Example of data that does not require un update: PPGF

PPGR = (PPGIO + PPGAO) * PPGF 
PPGF = CLIP(PPGF2,PPGF1,PYEAR) 
PPGF1 = PPGF2 = 1

PPGIO (resp. PPGAO) is the fraction of persistent pollutants generated from industrial (resp. agricultural) output. PPGF is a pollution generation factor; it is set by default to 1 and aims at simulating policies (e.g., reducing PPGF below 1 corresponds to introducing pollution abatement procedures that do not decrease resource consumption).

Example of parameters that require un update:

PPGIO = PCRUM * POP * FRPM * IMEF * IMTI 
FRPM = 0.02 
IMEF = 0.1 
IMTI = 10 
AMTI = 1

FRPM is the fraction of resources in the form of persistent materials. It is set to 2% [Census 1971, p. 656] by dividing the U.S. refinery net production of petrochemical feedstocks by the U.S. refinery net output.

IMEF is the fraction of persistent materials mobilised by industrial activities that eventually enter the environment (excluding materials employed in agriculture). It is set to 10% by an estimation of the authors.

IMTI is the industrial materials toxicity index. It set to 10, ten times the AMTI (the agriculture materials toxicity index).

The AMTI is set to 1, ten times less than the IMTI.

Example of time series that require un update:

POP = {(1900, 16.0); (1920,19.0); (1940,22.0), (1960,31.0); (1980,42.0); (2000,53.0); (2020,67.0); (2040,86.0); (2060,109.0); (2080,139.0); (2100,176.0)}

POP is the population number.

Getting modern data a) Some of the data are easy to retrieve. For instance, population data (including future projections based on the UN's medium-fertility scenario) is available here.

b) Some of the data requires aggregating data arising from different sources or a consensual definition on how to estimate it (in particular, when some data is not available).

• For example, for the parameter FRPM described above, if we want to stick to the way the authors estimated the parameter, we can get modern data on the U.S. refinery net production of petrochemical feedstocks (available here) and then we need to sum all of the U.S. refinery net production of other materials (from here) and making sure that all materials have been included.
• Moreover, we could also seek to estimate the parameter at the globe level (instead of restricting to the U.S.) or consider a different way to estimate the amount of persistent pollutants which was not available at that time but might be available today (see the following excerpt).

Overall, the process of updating with modern data requires:

• fixing one model (we described above World3, but we could start by a simpler model such as World2),
• grasping the definition of each parameters set up by the authors,
• establishing a consensual definition of the parameters depending on the modern data available,
• retrieving data from modern data catalogs (e.g., OurWorldInData, Statista, OECD Data, Nature Scientific Data, World Bank Open Data , UN Stats Open SDG Data Hub, etc) and carefully checking the sources. A suitable interface with the data catalogs would be useful.
• feeding data to the model.