Sei sulla pagina 1di 1

Modelling and Measurements of Atmospheric Carbon

CSIR Centre for Mathematical Modelling and Computer Simulation, Bangalore


Team :
Dr. N. K. Indira
Indira@cmmacs.ernet.in

Dr. P.S. Swathi


swathi@cmmacs.ernet.in

Prof. V.K. Gaur


gaur@cmmacs.ernet.in

Atmospheric transport model

Background A Global Scenario

Forward Method
Divide the earth into 11 land and 11 ocean regions. From each region emit a source of 1GTC/yr.

In the decade of 1990 the average fossil fuel carbon emission to the was
6.3 GTC/yr. In 2008 it was 8.5 GTC/yr. (Le Quere, Nature 2009)
The rate increase in the atmospheric CO2 has been 3.2 - 4.1 GTC/yr
Remaining carbon (roughly half the emission)- where does it go?
Many studies in the past few years have given different estimates of
sources and sinks. Need reliable estimates for CO2 mitigation.

Source

Land use

Forward mode

Net Primary Productivity


The model concentrations at the measurement stations were taken as the basis functions for inversions
Atmospheric
transport
H

CO2 net fluxes

a priori estimation

Budget
2000-2006
(GtC / year)

xb ; var/cov P

Global CO2 distribution modelled with MOZART

Atmos. Conc.
simulated
ymodel

Observations
yo ; var/cov R

Background CO2 simulations from 4 given sources

+7.6

Seasonal net ecosystem productivity

Inverse mode

+1.5

Land

Ocean

Temperate Asia

Fossil fuel emissions of 1990 and 1995

Estimated oceanic uptake (Takahashi, 2002)

Atmosphere -4.1
Sink

Flux CO2 (Gt C / yr)

Fossil fuel

This emission was held spatially uniform over the ocean while over land it was scaled in proportion to the

-2.8

Computation of CO2 distribution from a pulse emission for a month for each calendar month from 11 land

-2.2

and 11 ocean regions


The basis functions from these pulse-runs are used to calculate CO2 fluxes from the differences

GlobalView - established by NOAA and CMDL to measure CO2


concentrations at several locations around the world
TransCom - an experiment designed to compare model atmospheric
transport of carbon and subsequent inversion to obtain estimates of sources
and sinks by reconciling modelled and measured CO2 concentrations.
Lack of robustness due to poor coverage of measurement stations.

between observed CO2 concentrations and modelled background fluxes.


The resolution of the transport model is T42 which is 2.8 degrees x 2.8 degrees.
It has 28 vertical levels. The model uses NCEP 1996 winds to transport species.

Measurement and Analysis

Network Design

Inversion Method

A-priori CO2 flux


distribution
1 deg x 1 deg

The model for inversion is


Gm = d

Caribou instrument
Seasonal variations at Hanle
Pondicherry
Automatic Weather station

Flask measurement
et al., 1997

G is the model basis function,


m is the source/sink in 22 regions and
d is the mismatch between the measured and modelled CO2
concentrations

There are very few trace gases monitoring sites located in India and
Central Asia, and consequently it is difficult to make a comparison of
regional gradients and get reliable flux estimates. We have established
two new stations, a continuous one at Hanle (32.78N, 78.96E, 4517m
ASL) and a discrete flask at Pondicherry (12.012N, 79.89E, 20m ASL).
The Hanle station is a high-precision NDIR (Caribou) analyser with very
precise temperature and flow regulation, accurate calibration and a
reproducibility of 0.02 ppm.

Atmospheric composition measurements of useful quality are


expensive and difficult. Either they require development and
deployment of expensive instruments or the painstaking
analysis of flasks returned to central measurement facilities. It
would be helpful if augmentation of the network was guided
so as to produce maximal return. Quantitative network design
(QND) ] is the process of optimising observing networks for
their information content. We have applied the concepts of
QND to optimise CO2 observing systems for determining

The differences between the coastal site of Cape Rama (near Goa) and other sites are more
significant for CH4 and CO. For both species the Cape Rama data set shows a seasonal cycle
with amplitude of about 150 ppb, much higher than in Hanle and Mt Waliguan (China). The
observed gradients can be explained by the monsoon circulation. During the Indian summer
monsoon (June to September) winds from the southwest bring marine air mass to Cape Rama
while at Hanle the air mass has travelled over western India inducing a maximum level of CH4
concentration in August. No similar maximum is observed for CO. In winter the circulation
High uncertainties in Temperate
Asia due to sparse data

The inversions show offset from NEP especially


during the growing months indicating the presence
of terrestrial sinks in all the regions shown here.

Future measurement stations

patterns is reversed and maximum levels of CH4 and CO are observed at Cape Rama, as a
result of the regional Indian emissions.

The comparison of measurements at Hanle and Pondicherry shows


interesting features. For almost all the species the concentrations observed
at Pondicherry are higher than the ones observed at Hanle. This positive
offset at Pondicherry is related to the contribution of regional emissions
in South East India. The 100 ppb offset seen in CO is due to the important
contribution of biomass in India for cooking and heating.

The typical

CO2

diurnal

cycle is

characterized by amplitude of about 0.5


ppm. It is maximum in Winter/Fall, with
highest CO2 concentration observed
during the afternoon. The CO2 daily
maximum corresponds to a period of
increase of the wind speed from 2-3
m/sec up to 7-10 m/sec. At the same
time there is also a shift in wind

3 days - backtrajectories (Hysplit model)

direction from 160 to 210. In summer


the CO2 diurnal cycle presents a first
plateau in afternoon, and a highest one
between 8pm and midnight. Such CO2

July 2005

variability is clearly linked to the local


meteorology and which implies that the
CO2 concentrations is different from
daytime to night time. This is a typical
feature that is

observed at mountain

stations. However, due to the low


density of CO2 local sources around
Hanle observatory, the difference of

January 2005

concentration remains very low and


consequently

all

valid

data

are

considered as representative of large


scale surface fluxes

surface sources and sinks

Potrebbero piacerti anche