Thank you for so thoughtfully pointing out my oversight that 6.6 PgC is
not the same as 6.6 PgC of CO2.  So if Carbon is not the same as carbon
dioxide then just how much of the 6.6 PgC actually becomes carbon
dioxide. Your calculation assumes that all of that carbon from
industrial activity becomes CO2. Is that realistic? How much of that
carbon is in the form of particulates and carbon compounds other than CO2?

What is most interesting about the study by Grubler is the admission
that they really don't know how much CO2 is emitted. They have to
guesstimate based on economic activity. On page 13 he describes how they
calculate carbon emissions -
"... emissions, i.e. the efficiency of energy use per unit of GDP
(Energy/GDP) as well as the carbon intensity of energy used (carbon
dioxide/Energy)."

This may delight mathematicians but is it a reliable way to measure the
amount of  CO2 (not just carbon - which is what they seem to be
measuring) emitted from human activity?

Moreover, Grubler states that "human alterations to the carbon cycle are
comparatively small, are difficult to quantify and are hence subject to
considerable uncertainties, especially for land use change carbon
emissions." pg 2;

In fact, uncertainty seems to be the byword for this report.
Page 2 - "This estimated net biospheric flux is the difference between
estimates of a variety of carbon sources and sinks ... and is affected
by high uncertainty margins."
Page 5 - "Whereas bottom-up assessments thus continue to be the best way
for estimating current and historical gross carbon emissions, it is not
possible yet to accurately determine net biospheric emissions."
Page 15- "Because the energy use data include the historically important
non-commercial uses of traditional biofuels, the resulting energy use
and energy intensities ... estimates are subject to a high degree of
uncertainty; the values given here are represent a conservative lower
bound estimate ..."

Emissions are on thing, atmospheric concentrations are another. Grubler
states at page 2 "The inter-annual variations of measured atmospheric
carbon dioxide concentrations (5.4 parts per million by volume (ppmv) in
1998 at Muana Loa, Hawaii ... clearly indicate the importance of
seasonal vegetation cycles in biospheric carbon fluxes. The inter-annual
variation of 5.4 ppmv corresponds to some 11PgC.  Compared to the size
of annual fluxes that characterize the carbon cycle and its inter-annual
variations, anthropogenic alterations to the carbon cycle are
comparatively small and hence impossible to measure directly."

Your 15 billion tonnes per year is slightly more than the inter-annual
variation.
The point of all this is that man is not yet capable of measuring the
anthropogenic carbon (or CO2) emissions, and their best guesses show
that it is "comparatively small" compared to natural sources.

Let's not pretend that our numbers are derived with scientific certainty.

Chuck