Search Results
You are looking at 1 - 3 of 3 items for
- Author or Editor: Colin Morice x
- Refine by Access: All Content x
Abstract
Surface temperature is a vital metric of Earth’s climate state but is incompletely observed in both space and time: over half of monthly values are missing from the widely used HadCRUT4.6 global surface temperature dataset. Here we apply the graphical expectation–maximization algorithm (GraphEM), a recently developed imputation method, to construct a spatially complete estimate of HadCRUT4.6 temperatures. GraphEM leverages Gaussian Markov random fields (also known as Gaussian graphical models) to better estimate covariance relationships within a climate field, detecting anisotropic features such as land–ocean contrasts, orography, ocean currents, and wave-propagation pathways. This detection leads to improved estimates of missing values compared to methods (such as kriging) that assume isotropic covariance relationships, as we show with real and synthetic data. This interpolated analysis of HadCRUT4.6 data is available as a 100-member ensemble, propagating information about sampling variability available from the original HadCRUT4.6 dataset. A comparison of Niño-3.4 and global mean monthly temperature series with published datasets reveals similarities and differences due in part to the spatial interpolation method. Notably, the GraphEM-completed HadCRUT4.6 global temperature displays a stronger early twenty-first-century warming trend than its uninterpolated counterpart, consistent with recent analyses using other datasets. Known events like the 1877/78 El Niño are recovered with greater fidelity than with kriging, and result in different assessments of changes in ENSO variability through time. Gaussian Markov random fields provide a more geophysically motivated way to impute missing values in climate fields, and the associated graph provides a powerful tool to analyze the structure of teleconnection patterns. We close with a discussion of wider applications of Markov random fields in climate science.
Abstract
Surface temperature is a vital metric of Earth’s climate state but is incompletely observed in both space and time: over half of monthly values are missing from the widely used HadCRUT4.6 global surface temperature dataset. Here we apply the graphical expectation–maximization algorithm (GraphEM), a recently developed imputation method, to construct a spatially complete estimate of HadCRUT4.6 temperatures. GraphEM leverages Gaussian Markov random fields (also known as Gaussian graphical models) to better estimate covariance relationships within a climate field, detecting anisotropic features such as land–ocean contrasts, orography, ocean currents, and wave-propagation pathways. This detection leads to improved estimates of missing values compared to methods (such as kriging) that assume isotropic covariance relationships, as we show with real and synthetic data. This interpolated analysis of HadCRUT4.6 data is available as a 100-member ensemble, propagating information about sampling variability available from the original HadCRUT4.6 dataset. A comparison of Niño-3.4 and global mean monthly temperature series with published datasets reveals similarities and differences due in part to the spatial interpolation method. Notably, the GraphEM-completed HadCRUT4.6 global temperature displays a stronger early twenty-first-century warming trend than its uninterpolated counterpart, consistent with recent analyses using other datasets. Known events like the 1877/78 El Niño are recovered with greater fidelity than with kriging, and result in different assessments of changes in ENSO variability through time. Gaussian Markov random fields provide a more geophysically motivated way to impute missing values in climate fields, and the associated graph provides a powerful tool to analyze the structure of teleconnection patterns. We close with a discussion of wider applications of Markov random fields in climate science.
Abstract
The transient climate response (TCR) quantifies the warming expected during a transient doubling of greenhouse gas concentrations in the atmosphere. Many previous studies quantifying the observed historic response to greenhouse gases, and with it the TCR, use multimodel mean fingerprints and found reasonably constrained values, which contributed to the IPCC estimated (>66%) range from 1° to 2.5°C. Here, it is shown that while the multimodel mean fingerprint is statistically more powerful than any individual model’s fingerprint, it does lead to overconfident results when applied to synthetic data, if model uncertainty is neglected. Here, a Bayesian method is used that estimates TCR, accounting for climate model and observational uncertainty with indices of global temperature that aim at constraining the aerosol contribution to the historical record better. Model uncertainty in the aerosol response was found to be large. Nevertheless, an overall TCR estimate of 0.4°–3.1°C (>90%) was calculated from the historical record, which reduces to 1.0°–2.6°C when using prior information that rules out negative TCR values and model misestimates of more than a factor of 3, and to 1.2°–2.4°C when using the multimodel mean fingerprints with a variance correction. Modeled temperature, like in the observations, is calculated as a blend of sea surface and air temperatures.
Abstract
The transient climate response (TCR) quantifies the warming expected during a transient doubling of greenhouse gas concentrations in the atmosphere. Many previous studies quantifying the observed historic response to greenhouse gases, and with it the TCR, use multimodel mean fingerprints and found reasonably constrained values, which contributed to the IPCC estimated (>66%) range from 1° to 2.5°C. Here, it is shown that while the multimodel mean fingerprint is statistically more powerful than any individual model’s fingerprint, it does lead to overconfident results when applied to synthetic data, if model uncertainty is neglected. Here, a Bayesian method is used that estimates TCR, accounting for climate model and observational uncertainty with indices of global temperature that aim at constraining the aerosol contribution to the historical record better. Model uncertainty in the aerosol response was found to be large. Nevertheless, an overall TCR estimate of 0.4°–3.1°C (>90%) was calculated from the historical record, which reduces to 1.0°–2.6°C when using prior information that rules out negative TCR values and model misestimates of more than a factor of 3, and to 1.2°–2.4°C when using the multimodel mean fingerprints with a variance correction. Modeled temperature, like in the observations, is calculated as a blend of sea surface and air temperatures.
Abstract
Time series of global and regional mean surface air temperature (SAT) anomalies are a common metric used to estimate recent climate change. Various techniques can be used to create these time series from meteorological station data. The degree of difference arising from using five different techniques, based on existing temperature anomaly dataset techniques, to estimate Arctic SAT anomalies over land and sea ice was investigated using reanalysis data as a test bed. Techniques that interpolated anomalies were found to result in smaller errors than noninterpolating techniques relative to the reanalysis reference. Kriging techniques provided the smallest errors in estimates of Arctic anomalies, and simple kriging was often the best kriging method in this study, especially over sea ice. A linear interpolation technique had, on average, root-mean-square errors (RMSEs) up to 0.55 K larger than the two kriging techniques tested. Noninterpolating techniques provided the least representative anomaly estimates. Nonetheless, they serve as useful checks for confirming whether estimates from interpolating techniques are reasonable. The interaction of meteorological station coverage with estimation techniques between 1850 and 2011 was simulated using an ensemble dataset comprising repeated individual years (1979–2011). All techniques were found to have larger RMSEs for earlier station coverages. This supports calls for increased data sharing and data rescue, especially in sparsely observed regions such as the Arctic.
Abstract
Time series of global and regional mean surface air temperature (SAT) anomalies are a common metric used to estimate recent climate change. Various techniques can be used to create these time series from meteorological station data. The degree of difference arising from using five different techniques, based on existing temperature anomaly dataset techniques, to estimate Arctic SAT anomalies over land and sea ice was investigated using reanalysis data as a test bed. Techniques that interpolated anomalies were found to result in smaller errors than noninterpolating techniques relative to the reanalysis reference. Kriging techniques provided the smallest errors in estimates of Arctic anomalies, and simple kriging was often the best kriging method in this study, especially over sea ice. A linear interpolation technique had, on average, root-mean-square errors (RMSEs) up to 0.55 K larger than the two kriging techniques tested. Noninterpolating techniques provided the least representative anomaly estimates. Nonetheless, they serve as useful checks for confirming whether estimates from interpolating techniques are reasonable. The interaction of meteorological station coverage with estimation techniques between 1850 and 2011 was simulated using an ensemble dataset comprising repeated individual years (1979–2011). All techniques were found to have larger RMSEs for earlier station coverages. This supports calls for increased data sharing and data rescue, especially in sparsely observed regions such as the Arctic.
