Abstract
The tropical cloud forest ecosystem in western equatorial Africa (WEA) is known to be sensitive to the presence of an extensive and persistent low-level stratiform cloud deck during the long dry season from June to September (JJAS). Here, we present a new climatology of the diurnal cycle of the low-level cloud cover from surface synoptic stations over WEA during JJAS 1971–2019. For the period JJAS 2008–19, we also utilized estimates of cloudiness from four satellite products, namely, the Satellite Application Facility on Support to Nowcasting and Very Short Range Forecasting (SAFNWC) cloud classification, the Day and Night Microphysical Schemes (DMS/NMS), and cross sections from CALIPSO and CloudSat (2B-GEOPROF-lidar). A comparison with surface stations reveals that the NMS at night together with SAFNWC at daytime yield the smallest biases. The climatological analysis reveals that low-level clouds persist during the day over the coastal plains and windward side of the low mountain ranges. Conversely, on their leeward sides, i.e., over the plateaus, a decrease of the low-level cloud frequency is observed in the afternoon, together with a change from stratocumulus to cumulus. At night, the low-level cloud deck reforms over this region with the largest cloud occurrence frequencies in the morning. Vertical profiles from 2B-GEOPROF-lidar reveal cloud tops below 3000 m even at daytime. The station data and the suitable satellite products form the basis to better understand the physical processes controlling the clouds and to evaluate cloudiness from reanalyses and models.
Abstract
The tropical cloud forest ecosystem in western equatorial Africa (WEA) is known to be sensitive to the presence of an extensive and persistent low-level stratiform cloud deck during the long dry season from June to September (JJAS). Here, we present a new climatology of the diurnal cycle of the low-level cloud cover from surface synoptic stations over WEA during JJAS 1971–2019. For the period JJAS 2008–19, we also utilized estimates of cloudiness from four satellite products, namely, the Satellite Application Facility on Support to Nowcasting and Very Short Range Forecasting (SAFNWC) cloud classification, the Day and Night Microphysical Schemes (DMS/NMS), and cross sections from CALIPSO and CloudSat (2B-GEOPROF-lidar). A comparison with surface stations reveals that the NMS at night together with SAFNWC at daytime yield the smallest biases. The climatological analysis reveals that low-level clouds persist during the day over the coastal plains and windward side of the low mountain ranges. Conversely, on their leeward sides, i.e., over the plateaus, a decrease of the low-level cloud frequency is observed in the afternoon, together with a change from stratocumulus to cumulus. At night, the low-level cloud deck reforms over this region with the largest cloud occurrence frequencies in the morning. Vertical profiles from 2B-GEOPROF-lidar reveal cloud tops below 3000 m even at daytime. The station data and the suitable satellite products form the basis to better understand the physical processes controlling the clouds and to evaluate cloudiness from reanalyses and models.
Abstract
In this study, we design a statistical method to couple observations with a physics-based tropical cyclone (TC) rainfall model (TCR) and engineered-synthetic storms for assessing TC rainfall hazard. We first propose a bias-correction method to minimize the errors induced by TCR via matching the probability distribution of TCR-simulated historical TC rainfall with gauge observations. Then we assign occurrence probabilities to engineered-synthetic storms to reflect local climatology, through a resampling method that matches the probability distribution of a newly-proposed storm parameter named rainfall potential (POT) in the synthetic dataset with that in the observation. POT is constructed to include several important storm parameters for TC rainfall such as TC intensity, duration, and distance and environmental humidity near landfall, and it is shown to be correlated with TCR-simulated rainfall. The proposed method has a satisfactory performance in reproducing the rainfall hazard curve in various locations in continental U. S.; it is an improvement over the traditional joint probability method (JPM) for TC rainfall hazard assessment.
Abstract
In this study, we design a statistical method to couple observations with a physics-based tropical cyclone (TC) rainfall model (TCR) and engineered-synthetic storms for assessing TC rainfall hazard. We first propose a bias-correction method to minimize the errors induced by TCR via matching the probability distribution of TCR-simulated historical TC rainfall with gauge observations. Then we assign occurrence probabilities to engineered-synthetic storms to reflect local climatology, through a resampling method that matches the probability distribution of a newly-proposed storm parameter named rainfall potential (POT) in the synthetic dataset with that in the observation. POT is constructed to include several important storm parameters for TC rainfall such as TC intensity, duration, and distance and environmental humidity near landfall, and it is shown to be correlated with TCR-simulated rainfall. The proposed method has a satisfactory performance in reproducing the rainfall hazard curve in various locations in continental U. S.; it is an improvement over the traditional joint probability method (JPM) for TC rainfall hazard assessment.
Abstract
Many studies have aimed to identify novel storm characteristics that are indicative of current or future severe weather potential using a combination of ground-based radar observations and severe reports. However, this is often done on a small scale using limited case studies on the order of tens to hundreds of storms due to how time-intensive this process is. Herein, we introduce the GridRad-Severe dataset, a database including ∼100 severe weather days per year and upwards of 1.3 million objectively tracked storms from 2010-2019. Composite radar volumes spanning objectively determined, report-centered domains are created for each selected day using the GridRad compositing technique, with dates objectively determined using report thresholds defined to capture the highest-end severe weather days from each year, evenly distributed across all severe report types (tornadoes, severe hail, and severe wind). Spatiotemporal domain bounds for each event are objectively determined to encompass both the majority of reports as well as the time of convection initiation. Severe weather reports are matched to storms that are objectively tracked using the radar data, so the evolution of the storm cells and their severe weather production can be evaluated. Herein, we apply storm mode (single cell, multicell, or mesoscale convective system) and right-moving supercell classification techniques to the dataset, and revisit various questions about severe storms and their bulk characteristics posed and evaluated in past work. Additional applications of this dataset are reviewed for possible future studies.
Abstract
Many studies have aimed to identify novel storm characteristics that are indicative of current or future severe weather potential using a combination of ground-based radar observations and severe reports. However, this is often done on a small scale using limited case studies on the order of tens to hundreds of storms due to how time-intensive this process is. Herein, we introduce the GridRad-Severe dataset, a database including ∼100 severe weather days per year and upwards of 1.3 million objectively tracked storms from 2010-2019. Composite radar volumes spanning objectively determined, report-centered domains are created for each selected day using the GridRad compositing technique, with dates objectively determined using report thresholds defined to capture the highest-end severe weather days from each year, evenly distributed across all severe report types (tornadoes, severe hail, and severe wind). Spatiotemporal domain bounds for each event are objectively determined to encompass both the majority of reports as well as the time of convection initiation. Severe weather reports are matched to storms that are objectively tracked using the radar data, so the evolution of the storm cells and their severe weather production can be evaluated. Herein, we apply storm mode (single cell, multicell, or mesoscale convective system) and right-moving supercell classification techniques to the dataset, and revisit various questions about severe storms and their bulk characteristics posed and evaluated in past work. Additional applications of this dataset are reviewed for possible future studies.
Abstract
Mesoscale eddies can alter the propagation of wind-generated near-inertial waves (NIWs). Different from previous studies, the subsurface mooring observed NIWs are generated outside an anticyclonic eddy (ACE) and then interact with the arriving ACE. It is found that with the arrival of the ACE, the NIWs accelerate to propagate downward and the maximum vertical wavelength and group velocity of NIWs reach ~500 m and ~35 m/day, respectively. When entering the core of the ACE, the near-inertial energy is trapped, and finally stalls at a critical depth, which basically corresponds to the base of the ACE located at around 750 m depth. Through a ray-tracing model and dynamic analyses, this critical depth is much deeper than that of NIWs generated directly inside an ACE. By using depth-time varying stratification and relative vorticity, ray-tracing experiments further demonstrate that NIWs generated outside and passed over by an ACE can propagate to deep depths. Furthermore, energy budget analyses indicate that the net energy transfer from the ACE to NIWs plays an important role in the enhancement of downward-propagating near-inertial energy and its long-term persistence (~45 days) in the critical layer. Within the critical layer, the enhancement of shear instability and nonlinear interactions among internal waves account for the loss of the trapped near-in ertial energy and provide energy for furnishing deep ocean mixing.
Abstract
Mesoscale eddies can alter the propagation of wind-generated near-inertial waves (NIWs). Different from previous studies, the subsurface mooring observed NIWs are generated outside an anticyclonic eddy (ACE) and then interact with the arriving ACE. It is found that with the arrival of the ACE, the NIWs accelerate to propagate downward and the maximum vertical wavelength and group velocity of NIWs reach ~500 m and ~35 m/day, respectively. When entering the core of the ACE, the near-inertial energy is trapped, and finally stalls at a critical depth, which basically corresponds to the base of the ACE located at around 750 m depth. Through a ray-tracing model and dynamic analyses, this critical depth is much deeper than that of NIWs generated directly inside an ACE. By using depth-time varying stratification and relative vorticity, ray-tracing experiments further demonstrate that NIWs generated outside and passed over by an ACE can propagate to deep depths. Furthermore, energy budget analyses indicate that the net energy transfer from the ACE to NIWs plays an important role in the enhancement of downward-propagating near-inertial energy and its long-term persistence (~45 days) in the critical layer. Within the critical layer, the enhancement of shear instability and nonlinear interactions among internal waves account for the loss of the trapped near-in ertial energy and provide energy for furnishing deep ocean mixing.
Abstract
We investigate the land–ocean warming contrast mechanisms, ϕ, defined as the land-mean surface air temperature (SAT) change divided by the ocean-mean SAT change, in a transient climate response (TCR) obtained from the Coupled Model Intercomparison Project phase 6 (CMIP6) 1% per year CO2 increase experiments (1pctCO2). The energy budget framework devised in Part I is applied to 15 CMIP6 1pctCO2 simulations, and the climate response in year 140 when the CO2 concentration was quadrupled was compared with a near-equilibrium climate response (NEQ), defined as the last 30-yr mean in the abrupt CO2 quadrupling (abrupt4×CO2) experiments. It is shown that ϕ is larger in TCR than in NEQ by approximately 4%, although the difference is not statistically significant. In TCR, effective radiative forcing is large over land compared to the ocean, and this is the main contributor to ϕ as in NEQ. The time evolution of ϕ in 1pctCO2 can be reconstructed by means of the fast and slow components of climate response in abrupt4×CO2, indicating that the essential mechanism for the land–ocean warming contrast shown in Part I applies to TCR. Furthermore, our analyses reveal a compensation between land-to-ocean atmospheric energy transport that decreases ϕ and ocean heat uptake that increases ϕ. Regardless of the time scale of the response, these two processes are linked by the change in atmospheric circulation, leading to the small combined effect. As a result, the multimodel mean ϕ in 1pctCO2 is roughly time invariant at approximately 1.5 despite the continuous increase in CO2.
Significance Statement
The land–ocean warming contrast, which indicates large land surface warming compared to ocean surface warming in response to an increase in atmospheric CO2 concentration, is a striking feature of human-induced global warming. This study focuses on temporal changes in the magnitude of the land–ocean warming contrast in transient climate change simulations and shows that the magnitude of the land–ocean warming contrast is nearly constant over time, maintaining a ratio of approximately 1.5, between land and ocean surface warming. This small temporal change is explained mainly by a compensation between land-to-ocean energy transport and ocean heat uptake, because both act in opposite directions to the land–ocean warming contrast.
Abstract
We investigate the land–ocean warming contrast mechanisms, ϕ, defined as the land-mean surface air temperature (SAT) change divided by the ocean-mean SAT change, in a transient climate response (TCR) obtained from the Coupled Model Intercomparison Project phase 6 (CMIP6) 1% per year CO2 increase experiments (1pctCO2). The energy budget framework devised in Part I is applied to 15 CMIP6 1pctCO2 simulations, and the climate response in year 140 when the CO2 concentration was quadrupled was compared with a near-equilibrium climate response (NEQ), defined as the last 30-yr mean in the abrupt CO2 quadrupling (abrupt4×CO2) experiments. It is shown that ϕ is larger in TCR than in NEQ by approximately 4%, although the difference is not statistically significant. In TCR, effective radiative forcing is large over land compared to the ocean, and this is the main contributor to ϕ as in NEQ. The time evolution of ϕ in 1pctCO2 can be reconstructed by means of the fast and slow components of climate response in abrupt4×CO2, indicating that the essential mechanism for the land–ocean warming contrast shown in Part I applies to TCR. Furthermore, our analyses reveal a compensation between land-to-ocean atmospheric energy transport that decreases ϕ and ocean heat uptake that increases ϕ. Regardless of the time scale of the response, these two processes are linked by the change in atmospheric circulation, leading to the small combined effect. As a result, the multimodel mean ϕ in 1pctCO2 is roughly time invariant at approximately 1.5 despite the continuous increase in CO2.
Significance Statement
The land–ocean warming contrast, which indicates large land surface warming compared to ocean surface warming in response to an increase in atmospheric CO2 concentration, is a striking feature of human-induced global warming. This study focuses on temporal changes in the magnitude of the land–ocean warming contrast in transient climate change simulations and shows that the magnitude of the land–ocean warming contrast is nearly constant over time, maintaining a ratio of approximately 1.5, between land and ocean surface warming. This small temporal change is explained mainly by a compensation between land-to-ocean energy transport and ocean heat uptake, because both act in opposite directions to the land–ocean warming contrast.
Abstract
Using observation and reanalysis data, we investigated the effect of the sea surface temperature anomalies associated with ENSO Modoki from September to October on interannual variations in Antarctic stratospheric ozone from October to November. It was found that the planetary wave anomalies generated by ENSO Modoki in the tropical troposphere propagate to the southern mid- and then high-latitude stratosphere. The planetary wave anomalies have a profound impact on the polar vortex, subsequently affecting the interannual variations in Antarctic stratospheric ozone. Further analysis revealed that the responses of the polar vortex and ozone to ENSO Modoki are mainly modulated by the wave-1 and wave-3 components, and the effect of wave-2 is opposite and offset by those of wave-1 and wave-3. The contribution of the residual waves (after removing waves 1, 2 and 3, and the remaining waves) are relatively small. Furthermore, we evaluated the performance of CMIP6 models in simulating the impacts of ENSO Modoki on the southern stratospheric polar vortex and ozone. We selected seven models, that include stratospheric processes and stratospheric chemical ozone. We found that all of them can capable of distinguishing between eastern Pacific ENSO and ENSO Modoki events. However, only GISS-E2-1-G and MPI-ESM-1-2-HAM can simulate the patterns of ozone, circulation and temperature in the Southern Hemisphere in a manner that closely resembles the reanalysis results. Further analysis indicated that these two models can better simulate the propagation of planetary wave activities in the troposphere forced by ENSO Modoki, whereas the other models produce significantly different results to those obtained from observations.
Abstract
Using observation and reanalysis data, we investigated the effect of the sea surface temperature anomalies associated with ENSO Modoki from September to October on interannual variations in Antarctic stratospheric ozone from October to November. It was found that the planetary wave anomalies generated by ENSO Modoki in the tropical troposphere propagate to the southern mid- and then high-latitude stratosphere. The planetary wave anomalies have a profound impact on the polar vortex, subsequently affecting the interannual variations in Antarctic stratospheric ozone. Further analysis revealed that the responses of the polar vortex and ozone to ENSO Modoki are mainly modulated by the wave-1 and wave-3 components, and the effect of wave-2 is opposite and offset by those of wave-1 and wave-3. The contribution of the residual waves (after removing waves 1, 2 and 3, and the remaining waves) are relatively small. Furthermore, we evaluated the performance of CMIP6 models in simulating the impacts of ENSO Modoki on the southern stratospheric polar vortex and ozone. We selected seven models, that include stratospheric processes and stratospheric chemical ozone. We found that all of them can capable of distinguishing between eastern Pacific ENSO and ENSO Modoki events. However, only GISS-E2-1-G and MPI-ESM-1-2-HAM can simulate the patterns of ozone, circulation and temperature in the Southern Hemisphere in a manner that closely resembles the reanalysis results. Further analysis indicated that these two models can better simulate the propagation of planetary wave activities in the troposphere forced by ENSO Modoki, whereas the other models produce significantly different results to those obtained from observations.
Abstract
The radiative forcing of carbon dioxide (CO2) at the top-of-atmosphere (TOA) has a rich spatial structure and has implications for large-scale climate changes, such as poleward energy transport and tropical circulation change. Beyond the TOA, additional CO2 increases downwelling longwave at the surface, and this change in flux is the surface CO2 forcing. Here, we thoroughly evaluate the spatiotemporal variation of the instantaneous, longwave CO2 radiative forcing at both the TOA and surface. The instantaneous forcing is calculated with a radiative transfer model using ERA5 reanalysis fields. Multivariate regression models show that the broadband forcing at the TOA and surface are well-predicted by local temperatures, humidity, and cloud radiative effects. The difference between the TOA and surface forcing, the atmospheric forcing, can be either positive or negative and is mostly controlled by the column water vapor, with little explicit dependence on the surface temperature. The role of local variables on the TOA forcing is also assessed by partitioning the change in radiative flux to the component emitted by the surface vs. that emitted by the atmosphere. In cold, dry regions, the surface and atmospheric contribution partially cancel out, leading to locally weak or even negative TOA forcing. In contrast, in the warm, moist regions, the surface and atmospheric components strengthen each other, resulting in overall larger TOA forcing. The relative contribution of surface and atmosphere to the TOA forcing depends on the optical thickness in the current climate, which, in turn, is controlled by the column water vapor.
Abstract
The radiative forcing of carbon dioxide (CO2) at the top-of-atmosphere (TOA) has a rich spatial structure and has implications for large-scale climate changes, such as poleward energy transport and tropical circulation change. Beyond the TOA, additional CO2 increases downwelling longwave at the surface, and this change in flux is the surface CO2 forcing. Here, we thoroughly evaluate the spatiotemporal variation of the instantaneous, longwave CO2 radiative forcing at both the TOA and surface. The instantaneous forcing is calculated with a radiative transfer model using ERA5 reanalysis fields. Multivariate regression models show that the broadband forcing at the TOA and surface are well-predicted by local temperatures, humidity, and cloud radiative effects. The difference between the TOA and surface forcing, the atmospheric forcing, can be either positive or negative and is mostly controlled by the column water vapor, with little explicit dependence on the surface temperature. The role of local variables on the TOA forcing is also assessed by partitioning the change in radiative flux to the component emitted by the surface vs. that emitted by the atmosphere. In cold, dry regions, the surface and atmospheric contribution partially cancel out, leading to locally weak or even negative TOA forcing. In contrast, in the warm, moist regions, the surface and atmospheric components strengthen each other, resulting in overall larger TOA forcing. The relative contribution of surface and atmosphere to the TOA forcing depends on the optical thickness in the current climate, which, in turn, is controlled by the column water vapor.
Abstract
The impact of North Indian atmospheric diabatic heating variation on summer rainfall over Central Asia (CA) at an interannual scale during 1960–2019 was investigated from thermal adaptation and water vapor transportation perspective. The results showed that more precipitation in southeastern CA is associated with the southward subtropical westerly jet (SWJ), caused by the ascending motion and weakened water vapor output on the south side. When the SWJ moves southward, the high-level water vapor transportation on the south side changes from outward (−1.9 × 106 kg s−1) to inward (0.6 × 106 kg s−1), and the positive anomalous relative vorticity advections by the basic westerly winds produce ascending anomalies over southeastern CA. The position change in the SWJ was mainly related to atmospheric diabatic heating over northern India (NI). The thermal vorticity adaptation caused by a weakened heating rate over NI leads to an anomalous upper-level cyclone over southeastern CA, and the associated cold temperature advection eventually cools the upper troposphere of southeastern CA and reduces the temperature gradient at mid-to-high latitudes, leading to the southward SWJ. Thermal adaptation of the circulation and temperature anomaly over southeastern CA to the NI thermal forcing were also verified by numerical experiments. Both the abnormal ascending motions and the weakened outward water vapor associated with the southward SWJ, caused by the weakened heating rate over NI, lead to more summer rainfall in southeastern CA. The changes in diabatic heating over NI are closely related to Indian Ocean SST. When the Indian Ocean SST is warmer, the south Asian summer monsoon weakens, causing less precipitation and, thus, a weakened heating rate over NI.
Significance Statement
This study established that the northern Indian atmospheric diabatic heating anomalies associated with Indian Ocean SST variation play an important role in influencing precipitation in central Asia (CA). The weakening of the atmospheric diabatic heating over the NI would not only cause an abnormal cyclone and cooling over southern CA through thermal adaptation but also lead to southward subtropical westerly jet (SWJ), ascending motions, and decreased outward water vapor on the south side in southeastern CA, eventually resulting in more precipitation in southeastern CA. The results emphasize the influence of tropical SST and atmospheric heat sources on midlatitude climate and are important for understanding summer precipitation change in southeastern CA.
Abstract
The impact of North Indian atmospheric diabatic heating variation on summer rainfall over Central Asia (CA) at an interannual scale during 1960–2019 was investigated from thermal adaptation and water vapor transportation perspective. The results showed that more precipitation in southeastern CA is associated with the southward subtropical westerly jet (SWJ), caused by the ascending motion and weakened water vapor output on the south side. When the SWJ moves southward, the high-level water vapor transportation on the south side changes from outward (−1.9 × 106 kg s−1) to inward (0.6 × 106 kg s−1), and the positive anomalous relative vorticity advections by the basic westerly winds produce ascending anomalies over southeastern CA. The position change in the SWJ was mainly related to atmospheric diabatic heating over northern India (NI). The thermal vorticity adaptation caused by a weakened heating rate over NI leads to an anomalous upper-level cyclone over southeastern CA, and the associated cold temperature advection eventually cools the upper troposphere of southeastern CA and reduces the temperature gradient at mid-to-high latitudes, leading to the southward SWJ. Thermal adaptation of the circulation and temperature anomaly over southeastern CA to the NI thermal forcing were also verified by numerical experiments. Both the abnormal ascending motions and the weakened outward water vapor associated with the southward SWJ, caused by the weakened heating rate over NI, lead to more summer rainfall in southeastern CA. The changes in diabatic heating over NI are closely related to Indian Ocean SST. When the Indian Ocean SST is warmer, the south Asian summer monsoon weakens, causing less precipitation and, thus, a weakened heating rate over NI.
Significance Statement
This study established that the northern Indian atmospheric diabatic heating anomalies associated with Indian Ocean SST variation play an important role in influencing precipitation in central Asia (CA). The weakening of the atmospheric diabatic heating over the NI would not only cause an abnormal cyclone and cooling over southern CA through thermal adaptation but also lead to southward subtropical westerly jet (SWJ), ascending motions, and decreased outward water vapor on the south side in southeastern CA, eventually resulting in more precipitation in southeastern CA. The results emphasize the influence of tropical SST and atmospheric heat sources on midlatitude climate and are important for understanding summer precipitation change in southeastern CA.
Abstract
The two-step U-Net model (TU-Net) contains a western North Pacific subtropical high (WNPSH) prediction model and a precipitation prediction model fed by the WNPSH predictions, oceanic heat content, and surface temperature. The data-driven forecast model provides improved 4-month lead predictions of the WNPSH and precipitation in the middle and lower reaches of the Yangtze River (MLYR), which has important implications for water resources management and precipitation-related disaster prevention in China. When compared with five state-of-the-art dynamical climate models including the Climate Forecast System of Nanjing University of Information Science and Technology (NUIST-CFS1.0) and four models participating in the North American Multi-Model Ensemble (NMME) project, the TU-Net produces comparable skills in forecasting 4-month lead geopotential height and winds at the 500- and 850-hPa levels. For the 4-month lead prediction of precipitation over the MLYR region, the TU-Net has the best correlation scores and mean latitude-weighted RMSE in each summer month and in boreal summer [June–August (JJA)], and pattern correlation coefficient scores are slightly lower than the dynamical models only in June and JJA. In addition, the results show that the constructed TU-Net is also superior to most of the dynamical models in predicting 2-m air temperature in the MLYR region at a 4-month lead. Thus, the deep learning-based TU-Net model can provide a rapid and inexpensive way to improve the seasonal prediction of summer precipitation and 2-m air temperature over the MLYR region.
Significance Statement
The purpose of this study is to examine the seasonal predictive skill of the western North Pacific subtropical high anomalies and summer rainfall anomalies over the middle and lower reaches of the Yangtze River region by means of deep learning methods. Our deep learning model provides a rapid and inexpensive way to improve the seasonal prediction of summer precipitation as well as 2-m air temperature. The work has important implications for water resources management and precipitation-related disaster prevention in China and can be extended in the future to predict other climate variables as well.
Abstract
The two-step U-Net model (TU-Net) contains a western North Pacific subtropical high (WNPSH) prediction model and a precipitation prediction model fed by the WNPSH predictions, oceanic heat content, and surface temperature. The data-driven forecast model provides improved 4-month lead predictions of the WNPSH and precipitation in the middle and lower reaches of the Yangtze River (MLYR), which has important implications for water resources management and precipitation-related disaster prevention in China. When compared with five state-of-the-art dynamical climate models including the Climate Forecast System of Nanjing University of Information Science and Technology (NUIST-CFS1.0) and four models participating in the North American Multi-Model Ensemble (NMME) project, the TU-Net produces comparable skills in forecasting 4-month lead geopotential height and winds at the 500- and 850-hPa levels. For the 4-month lead prediction of precipitation over the MLYR region, the TU-Net has the best correlation scores and mean latitude-weighted RMSE in each summer month and in boreal summer [June–August (JJA)], and pattern correlation coefficient scores are slightly lower than the dynamical models only in June and JJA. In addition, the results show that the constructed TU-Net is also superior to most of the dynamical models in predicting 2-m air temperature in the MLYR region at a 4-month lead. Thus, the deep learning-based TU-Net model can provide a rapid and inexpensive way to improve the seasonal prediction of summer precipitation and 2-m air temperature over the MLYR region.
Significance Statement
The purpose of this study is to examine the seasonal predictive skill of the western North Pacific subtropical high anomalies and summer rainfall anomalies over the middle and lower reaches of the Yangtze River region by means of deep learning methods. Our deep learning model provides a rapid and inexpensive way to improve the seasonal prediction of summer precipitation as well as 2-m air temperature. The work has important implications for water resources management and precipitation-related disaster prevention in China and can be extended in the future to predict other climate variables as well.
Abstract
On 7 February 2020 a relatively deep cyclone (~980 hPa) with mid-level frontogenesis produced heavy snow (20-30 mm liquid equivalent) over western and central New York State. Despite these characteristics, the precipitation was not organized into a narrow band of intensive snowfall. This event occurred during the Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS) field campaign. Using coordinated flight legs across New York State, a remote sensing aircraft (ER-2) sampled above the cloud while a P-3 aircraft collected in-cloud data. These data are used to validate several Weather Research and Forecasting (WRF) model simulations at 2-km and 0.67-km grid spacing using different initial and boundary conditions (RAP, GFS, and ERA5 analyses) and microphysics schemes (Thompson and P3). The differences between the WRF runs are used to explore sensitivity to initial conditions and microphysics schemes. All 18–24 h runs realistically produced a broad sloping region of frontogenesis at mid-levels typically; however, there were relatively large (20–30%) uncertainties in the magnitude of this forcing using different analyses and initialization times. The differences in surface precipitation distribution are small (< 10%) among the microphysics schemes, likely because there was little riming in the region of heaviest precipitation. Those runs with frontogenesis closest to the RAP analysis and a surface precipitation underprediction of 20–30% have too little ice aloft and at low-levels, suggesting deficiencies in ice generation and snow growth aloft in those runs. The 0.67-km grid produced more realistic convective cells aloft, but only 5–10% more precipitation than the 2-km grid.
Abstract
On 7 February 2020 a relatively deep cyclone (~980 hPa) with mid-level frontogenesis produced heavy snow (20-30 mm liquid equivalent) over western and central New York State. Despite these characteristics, the precipitation was not organized into a narrow band of intensive snowfall. This event occurred during the Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS) field campaign. Using coordinated flight legs across New York State, a remote sensing aircraft (ER-2) sampled above the cloud while a P-3 aircraft collected in-cloud data. These data are used to validate several Weather Research and Forecasting (WRF) model simulations at 2-km and 0.67-km grid spacing using different initial and boundary conditions (RAP, GFS, and ERA5 analyses) and microphysics schemes (Thompson and P3). The differences between the WRF runs are used to explore sensitivity to initial conditions and microphysics schemes. All 18–24 h runs realistically produced a broad sloping region of frontogenesis at mid-levels typically; however, there were relatively large (20–30%) uncertainties in the magnitude of this forcing using different analyses and initialization times. The differences in surface precipitation distribution are small (< 10%) among the microphysics schemes, likely because there was little riming in the region of heaviest precipitation. Those runs with frontogenesis closest to the RAP analysis and a surface precipitation underprediction of 20–30% have too little ice aloft and at low-levels, suggesting deficiencies in ice generation and snow growth aloft in those runs. The 0.67-km grid produced more realistic convective cells aloft, but only 5–10% more precipitation than the 2-km grid.
