“…Cross-validation is effective and widely used for assessing model prediction [24], and it can be used to compare and assess prediction accuracies between spatial-only and spacetime kriging [20,21,25]. As above, we considered a set of data denoted by the variable { ( , ) | , } Z Z s t s S t T that varies within a spatial domain S and time interval T. We let Z be observed at space-time points (s i ,t i ), i=1,···, n. Crossvalidation proceeds by removing an original observation datum Z(s j , t j ) and then making its prediction between these two data sets, mean absolute prediction error (MAE), and root mean square error (RMSE).…”
Section: Evaluation Of Methodsmentioning
confidence: 99%
“…where [19][20][21] and an extra global nugget N ST to capture the nugget effect [13], and is given by…”
Section: S T S T N H H S T H H R S T R S H T H N H Hmentioning
Observations of atmospheric carbon dioxide (CO 2 ) from satellites offer new data sources to understand global carbon cycling. The correlation structure of satellite-observed CO 2 can be analyzed and modeled by geostatistical methods, and CO 2 values at unsampled locations can be predicted with a correlation model. Conventional geostatistical analysis only investigates the spatial correlation of CO 2 , and does not consider temporal variation in the satellite-observed CO 2 data. In this paper, a spatiotemporal geostatistical method that incorporates temporal variability is implemented and assessed for analyzing the spatiotemporal correlation structure and prediction of monthly CO 2 in China. The spatiotemporal correlation is estimated and modeled by a product-sum variogram model with a global nugget component. The variogram result indicates a significant degree of temporal correlation within satellite-observed CO 2 data sets in China. Prediction of monthly CO 2 using the spatiotemporal variogram model and spacetime kriging procedure is implemented. The prediction is compared with a spatial-only geostatistical prediction approach using a cross-validation technique. The spatiotemporal approach gives better results, with higher correlation coefficient (r 2 ), and less mean absolute prediction error and root mean square error. Moreover, the monthly mapping result generated from the spatiotemporal approach has less prediction uncertainty and more detailed spatial variation of CO 2 than those from the spatial-only approach.CO 2 , Greenhouse Gases Observing Satellite (GOSAT), geostatistical analysis, space-time kriging, product-sum model
Citation:Zeng Z C, Lei L P, Guo L J, et al. Incorporating temporal variability to improve geostatistical analysis of satellite-observed
“…Cross-validation is effective and widely used for assessing model prediction [24], and it can be used to compare and assess prediction accuracies between spatial-only and spacetime kriging [20,21,25]. As above, we considered a set of data denoted by the variable { ( , ) | , } Z Z s t s S t T that varies within a spatial domain S and time interval T. We let Z be observed at space-time points (s i ,t i ), i=1,···, n. Crossvalidation proceeds by removing an original observation datum Z(s j , t j ) and then making its prediction between these two data sets, mean absolute prediction error (MAE), and root mean square error (RMSE).…”
Section: Evaluation Of Methodsmentioning
confidence: 99%
“…where [19][20][21] and an extra global nugget N ST to capture the nugget effect [13], and is given by…”
Section: S T S T N H H S T H H R S T R S H T H N H Hmentioning
Observations of atmospheric carbon dioxide (CO 2 ) from satellites offer new data sources to understand global carbon cycling. The correlation structure of satellite-observed CO 2 can be analyzed and modeled by geostatistical methods, and CO 2 values at unsampled locations can be predicted with a correlation model. Conventional geostatistical analysis only investigates the spatial correlation of CO 2 , and does not consider temporal variation in the satellite-observed CO 2 data. In this paper, a spatiotemporal geostatistical method that incorporates temporal variability is implemented and assessed for analyzing the spatiotemporal correlation structure and prediction of monthly CO 2 in China. The spatiotemporal correlation is estimated and modeled by a product-sum variogram model with a global nugget component. The variogram result indicates a significant degree of temporal correlation within satellite-observed CO 2 data sets in China. Prediction of monthly CO 2 using the spatiotemporal variogram model and spacetime kriging procedure is implemented. The prediction is compared with a spatial-only geostatistical prediction approach using a cross-validation technique. The spatiotemporal approach gives better results, with higher correlation coefficient (r 2 ), and less mean absolute prediction error and root mean square error. Moreover, the monthly mapping result generated from the spatiotemporal approach has less prediction uncertainty and more detailed spatial variation of CO 2 than those from the spatial-only approach.CO 2 , Greenhouse Gases Observing Satellite (GOSAT), geostatistical analysis, space-time kriging, product-sum model
Citation:Zeng Z C, Lei L P, Guo L J, et al. Incorporating temporal variability to improve geostatistical analysis of satellite-observed
“…Some applications are found in De Iaco et al (2002b), Li et al (2009), andDe Iaco (2010). This non-separable family of space-time covariances has been built by applying the convexity property of the covariances family.…”
Section: Product-sums and The Variogram Formmentioning
Positive definiteness represents an admissibility condition for a function to be a covariance. Nevertheless, the more restricted condition of strict positive definiteness has received attention in literature, especially in spatial statistics, since it ensures that the kriging system has a unique solution. Most known covariance functions are isotropic but there are applications where isotropy is not appropriate, e.g., spacetime covariance functions. One way to construct non-isotropic covariance functions is to use a product or a product-sum. In this article, it is given a necessary as well as a sufficient condition for a product of two covariance functions to be strictly positive definite. This result is extended to the well-known product-sum covariance model.
“…To estimate the model parameters, the iterative nonlinear weighted least-squared technique [37], [38] has been used. This procedure gives an estimate for the parameter vector ϕ, over all the possible values by minimizing the following sum:…”
A precise and high-resolution spatiotemporal distribution of atmospheric carbon dioxide (CO 2 ) is important in identifying and quantifying the CO 2 source and sinks on regional scales and emissions from discrete point sources. We propose the use of a regional gap-filling method by modeling the spatiotemporal correlation structures of column-averaged CO 2 dry air mole fractions (Xco 2 ) on a regional scale, using data from the Atmospheric CO 2 Observations from Space retrievals of the Greenhouse Gases Observing Satellite (ACOS-GOSAT) measurements over mainland China. The accuracy of the gap-filling results is verified by cross-validation and comparison with ground-based measurements. As the results of the spatiotemporal gap-filling method are applied to mainland China, the correlation coefficient (r 2 ) between the predicted values and true ones is greater than 0.85, the mean absolute prediction error is less than 1.5 ppm in cross-validation, and the seasonal cycle of the gap-filled data is generally in agreement with ground-based measurements. Finally, we compare the prediction accuracy based on our method with that based on the commonly used spatial-only kriging to further demonstrate the improved prediction accuracy. The applied regional gap-filling method, which makes full use of the multitemporal ACOS-GOSAT data, can generate a regional regular spatial distribution map of (Xco 2 ) at high spatial and temporal resolutions.Index Terms-Atmospheric CO 2 Observations from Space retrievals of the Greenhouse Gases Observing Satellite (ACOS-GOSAT), product-sum model, space-time kriging, Xco 2 .
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.