Networked environmental monitoring system and method

1. An environment monitoring method, comprising:

obtaining environmental information by combining a plurality of observations based on a plurality of sensors;

wherein the plurality of sensors include a first set of sensors each of a first accuracy at a first cost, and a second set of sensors each of a second accuracy at a second cost, the second accuracy being higher than the first accuracy, and the second cost being higher than the first cost, and

wherein the first set of sensors form a network comprising a plurality of distributed sensors at a first density;

the method further comprising:

obtaining environmental data with the plurality of distributed sensors, wherein the plurality of distributed sensors are calibrated to achieve a third accuracy suitable for environmental monitoring, and wherein the third accuracy is higher than the first accuracy;

assimilating the obtained environmental data together with meteorological information to derive information on the environment with a first resolution higher than a second resolution of information obtained with a network at a density lower than the first density;

calibrating the obtained environmental data by environmental correction through a successive regressing by solving:

y=a 0 x 0 +a 1 x 1 +a 2 x 2 + . . . a n x n +b+ε n

wherein y represents reported CO 2 value from the low-cost sensors, x 0 represents true CO 2 value, x 1 , x 2 , . . . , x n represent a number of other factors such as air pressure, temperature and humidity respectively;

the residuals ε n-1 and ε n at two successive regression steps are related to environmental variables using linear regression successively as:

ε n-1 =ε n −a n x n −b n ,

wherein n=1,3 for each environmental variable pressure p, temperature T, and water vapor q, wherein the linear regression method leads to eight correction coefficients, of the form a n and b n ,

the method further comprising:

applying the correction coefficients to the equation of y along with the environmental variables to correct sensor CO 2 observations for environmental influences:

y*={y−b 0 −( a 1 x 1 +b 1 ) . . . −( a n x n +b n )}/ a 0 ;

and

displaying the derived information on the environment on a display screen.

2. The method of claim 1 , wherein second set of sensors are configured for direct measurement and calibrating the first set of sensors; and

at least some of the plurality of distributed sensors are co-located with standard monitoring stations.

3. The method of claim 1 , further comprising an in-situ zero-drift correction.

4. The method of claim 1 , further comprising a network-enabled calibration.

5. An environment monitoring method, comprising:

obtaining environmental information by combining a plurality of observations based on a plurality of sensors;

wherein the plurality of sensors include a first set of sensors each of a first accuracy at a first cost, and a second set of sensors each of a second accuracy at a second cost, the second accuracy being higher than the first accuracy, and the second cost being higher than the first cost, and

wherein the first set of sensors form a network comprising a plurality of distributed sensors at a first density;

the method further comprising:

obtaining environmental data with the plurality of distributed sensors, wherein the plurality of distributed sensors are calibrated to achieve a third accuracy suitable for environmental monitoring, and wherein the third accuracy is higher than the first accuracy;

assimilating the obtained environmental data together with meteorological information to derive information on the environment with a first resolution higher than a second resolution of information obtained with a network at a density lower than the first density; and

a computational data assimilation to invert fluxes at a third spatiotemporal resolution to combine measured gas concentrations and meteorological information to derive information on pollution sources and sinks at the first resolution, including:

obtaining inputs from observations y o , the ensemble forecast

x k b ( t )= M ( x k a ( t− 1))

with mean x b and forecast of the observation

y k b =h ( x k b ) ,

wherein M represents a full nonlinear model, k is an index for model ensemble member, h is an observation operator mapping model prediction onto observation space to compute observation model error covariance y o −h(x k b );

applying the covariance as an ensemble square-root filter in which observations are assimilated to update only an ensemble mean while ensemble perturbations x k b − x b are updated by transforming forecast ensemble perturbations through a transform matrix:

x a = x b +X b {tilde over (P)} a ( HX b ) T R −1 [y o −h ( x b )]

X a =X b [( K− 1) {tilde over (P)} a ] 1/2 ,

wherein K is total number of ensemble members, X a , X b are perturbation matrices whose columns are analysis and the forecast ensemble perturbations, respectively. X b is updated every analysis time step, therefore forecast error covariance

P

b

=

1

K

-

1

⁢

X

b

⁢

X

b

T

is flow-dependent;

{tilde over (P)} a , an analysis error covariance in ensemble space, is given by

{tilde over (P)} a =[( K− 1) I +( HX b ) T R −1 ( HX b )] −1

which has dimension K by K, smaller than both dimension of the full non-linear model and number of observations;

thereby performing matrix inverse in the ensemble space spanned by forecast ensemble members, the reducing computational cost;

the method further comprising displaying the derived information on the environment over a map on a display screen.

6. The method of claim 5 , further comprising a vertical localization of column mixed CO 2 observations.

7. The method of claim 5 , further comprising a 4D assimilation.

8. The method of claim 5 , further comprising applying a short assimilation window.

9. The method of claim 5 , further comprising a temporal smoother.

10. The method of claim 5 , further comprising a time filter that combines inverted fluxes from previous several steps of data assimilation analysis as a weighted average.

11. An environment monitoring system, comprising:

a plurality of distributed sensors at a first density configured to obtain environmental data, wherein the plurality of distributed sensors comprise a first set of sensors each of a first accuracy at a first cost, and are calibrated to achieve a third accuracy higher than the first accuracy and suitable for environmental monitoring; and

one or more processing circuits configured to assimilate the obtained environmental data together with meteorological information to derive information on the environment with a first resolution higher than a second resolution of information obtained with a network at a density lower than the first density;

wherein the plurality of sensors include a first set of sensors each of a first accuracy at a first cost, and a second set of sensors each of a second accuracy at a second cost, the second accuracy being higher than the first accuracy, and the second cost being higher than the first cost;

the one or more processing circuits are further configured to:

obtain environmental data with the plurality of distributed sensors, wherein the plurality of distributed sensors are calibrated to achieve a third accuracy suitable for environmental monitoring, and wherein the third accuracy is higher than the first accuracy; and

calibrate the obtained environmental data by environmental correction through a multivariate regression to calculate regression coefficients for each of a plurality of K30 sensors to obtain five correction coefficients a n and b n , wherein n represents each independent variable, for dry CO 2 from Los Gatos Fast Greenhouse Gas Analyzers (LGRs), pressure P, temperature T, and water vapor mixing ratio q for an equation:

y*={y−b 0 −( a 1 x 1 +b 1 ) . . . −( a n x n +b n )}/ a 0

along with original K30 data, y, and environmental variables to predict true CO 2 concentration observed;

the system further comprising one or more computers to display the true CO 2 concentration observed.

12. The system of claim 11 , further comprising a base station configured to receive data from at least a subset of the plurality of distributed sensors.

13. The system of claim 11 , wherein the one or more processing circuits are further configured to manage the plurality of distributed sensors through at least one of Ethernet, cellular, or Wi-Fi communication channels.

14. The system of claim 11 , wherein:

the second set of sensors are configured for direct measurement and calibrating the first set of sensors; and

at least some of the plurality of distributed sensors are co-located with standard monitoring stations.