Climate Dynamics Spring 2015
All seminars Wednesdays 1:30 unless otherwise noted.
6 May: Natalie Burls, George Mason University
Exploratory Hall rm 3301
Deep water formation in the North Pacific during the warm Pliocene
The Pliocene Epoch, ~3-5 million years ago, is a particularly interesting period in Earth’s history. With a continental configuration similar to present-day, it is the most recent period within Earth’s history during which atmospheric CO2 levels are estimated to have been as high as today’s anthropogenically-forced levels. Our current understanding of Pliocene climate based on proxy evidence and modeling studies will be reviewed. New observational and modeling evidence will be presented suggesting that warm Pliocene conditions supported deep-water formation and meridional overturning during the Pliocene.
28 April: Andrew Badger, George Mason University
Tuesday, 9:00 am, 163 Research Hall
Thesis Defense:The Role of Large-Scale Land-Use Change on the Global Climate – Response and Sensitivity to Amazon Deforestation
Committee:P. Dirmeyer (dissertation director), Z. Boybeyi, T. DelSole, P. R. Houser
Large-scale land-use change, such as Amazon deforestation, can have a significant local effect on the climate and has the potential to impact the global climate system. Previous modeling studies have shown non-local responses due to Amazon deforestation. However, a common flaw in these studies is using prescribed ocean conditions, which can dampen the global response. This study uses a set of fully coupled modeling simulations to determine the responses and sensitivities to Amazon deforestation, both locally and globally. In addition, a set of realistic tropical crop vegetation types are developed for the Community Land Model version 4.5. The local increase in surface temperature and decrease in precipitation from this study are consistent with previous modeling studies. After deforestation, it was determined that stronger regions of land-atmospheric coupling are found in the formerly densely forested regions, while areas that receive irrigation become less coupled. This study highlights large-scale changes to the zonal and meridional circulation that are found to have impacts in remote regions throughout the tropics. Lastly, using a set of partial deforestation simulations, areas of non-linear responses to deforestation are found. A metric to quantify the degree of non-linearity over the spatial domain was developed; Amazon deforestation is found to have a tipping point effect for the climate system with less than half of the impacts within the spatial domain of influence being provided by 50% deforestation.
27 April: H. Annamalai, U. Hawaii
Monday, 11am, 3301 Exploratory Hall
Systematic errors in monsoon simulation: a way forward
In climate models, simulating the monsoon precipitation climatology remains a grand challenge. Compared to CMIP3, the multi-model-mean (MMM) errors for Asian-Australian monsoon (AAM) precipitation climatology in CMIP5, relative to GPCP observations, have shown little improvement. One of the implications is that uncertainties in the future projections of time-mean changes to AAM rainfall may not have reduced from CMIP3 to CMIP5. Despite dedicated efforts by the modeling community, the progress in monsoon modeling is rather slow. This leads us to wonder: Has the scientific community reached a “plateau” in modeling mean monsoon precipitation?
Our focus here is to improved understanding of the coupled air-sea interactions, and moist processes that govern the precipitation characteristics over the tropical Indian Ocean where large-scale errors persist. As a way forward, we propose three coordinated efforts, and they are: (i) idealized coupled model experiments; (ii) process-based diagnostics and (iii) direct observations to constrain model physics. We will argue that a systematic and coordinated approach in the identification of the various interactive processes that shape the precipitation basic state needs to be carried out, and high-quality observations over the data sparse monsoon region are needed to validate models and further improve model physics.
20 April: Eric Stofferahn, George Mason University
Monday, 9am, 256D/E Research Hall
Thesis Defense: Investigation of Aerosol Effects on the Arctic Surface Temperature during the Diurnal Cycle
Committee: Z. Boybeyi (dissertation director), J. Kinter, T. DelSole, M. Summers
Temperature changes in the Arctic due to anthropogenic climate change are larger in magnitude than those at lower latitudes, with sea ice extent and thickness diminishing since the dawn of the satellite era. Aerosols may play a vital role in determining the changes to the Arctic. Specifically, the ability of absorbing aerosols to change the vertical structure of the atmosphere and sulfate aerosols to act as cloud condensation nuclei play important parts in the maintenance of Arctic stratiform clouds. However, there are still large uncertainties in the impact of aerosols on the changes in Arctic surface temperature, particularly during the diurnal cycle. This study attempts to address these changes using the Weather, Research, and Forecasting Chemistry (WRF-CHEM) model. The study investigates the changes in surface temperature, as well as the variables which affect surface temperature, due to aerosol effects in the Arctic. A suite of ensemble runs are used to develop a filtering mechanism based upon the t-test to eliminate the effects of meteorological variability. The total aerosol effect is then separated into the changes caused by the aerosol direct effect, the aerosol semi-direct effect, and the aerosol indirect effects through the use of additional WRF-CHEM runs. The study shows that aerosol indirect effects are the dominant influence on surface temperature changes throughout the diurnal cycle. While much has been speculated about the cooling role of indirect aerosol effects, this study shows that the indirect effects have both a warming and cooling effect, depending upon the time of day, underlying surface properties, and aerosol size distribution/concentration.
25 March: Tiffany Smith, Johns Hopkins U.
3301 Exploratory Hall
Characterization, forcings, and feedbacks on summertime temperatures across the United States
Heat waves are the most deadly, and costly, natural disaster in the United States. As such, they are relevant to a wide range of stakeholders from climate scientists to public health experts. While there is value in the diversity of perspectives, this also leads to a lack of consensus heat wave definition, which can cause confusion while studying trends, processes, and impacts of heat wave events. First, this work will characterize spatial and temporal variations of heat wave occurrence across CONUS using multiple, previously published heat wave definitions. Next, summertime, daytime temperatures (Tmax) are used to regionalize CONUS through hierarchical clustering. Comprehensive analyses of the impacts global climate modes have on these regions Tmax are evaluated using a variety of statistical techniques. Given that these global-scale forcings cannot wholly explain the predictability of Tmax, this work will then discuss the role of regional land-atmosphere feedback mechanisms, specifically during the 2012 CONUS heat wave events. To do so, a series of sensitivity experiments will be presented that examine the timing, land surface initialization, and evolution of climate parameters.
23 March: Roop Saini
163 Research Hall, Monday, 1:30
Impact of Spring Soil Moisture Anomalies on Summer Precipitation over U.S. during Drought and Flood Events
This study examines the impact of soil moisture on summer (JJA) precipitation over U.S. during 1988 drought, 1993 flood and 2012 drought events using a regional climate model (RCM) RegCM4.1 coupled with Community Land Model Version 4 (CLM4). First, short-term sensitivity experiments are performed and necessary modifications are made to the moisture scheme to enhance the performance of the model in terms of precipitation. Second, using the modified moisture scheme, a 26-year (1987-2012) simulation driven with Reanalysis-2 lateral boundary conditions is conducted to evaluate the model performance with Climatic Research Unit (CRU) precipitation data and Reanalysis-2 low- level winds (850 hPa) data. The model performance is also compared with North American Regional Climate Change Assessment Programs (NARCCAP) ensemble of Reanalysis-driven RCMs, which shows RegCM4.1-CLM4 compares well with the NARCCAP models. The model is able to capture the extreme events of 1988 drought, 1993 flood, and 2012 drought. Third, spring and early summer soil moisture anomalies are examined for each extreme event with May and June showing the largest difference in the soil moisture anomalies. To investigate the response of summer precipitation to initial soil moisture conditions, May soil moisture values in the model are swapped between the flood and the drought years and the simulations starting May 1 are examined for 1988, 1993 and 2012. Swapping 1988 (drought) and 1993 (flood) May soil moisture values and swapping 2012 (drought) May soil moisture values with 1993 (flood) May soil moisture values show wet precipitation anomalies during 1988, dry precipitation anomalies during 1993, and wet precipitation anomalies during 2012 over Central U.S. The wet and dry precipitation anomalies during summer correspond well with the precipitation minus evaporation (P-E) anomalies, which is an indicator of the water resource conditions and relates to soil moisture availability. The impact of late spring soil moisture anomalies on subsequent precipitation is largest during JJA. This study reveals that late spring soil moisture is a useful predictor in summer precipitation.
16 March: Liang Chen, Texas A & M University
163 Research Hall, Monday, 10:30am
Climatic impacts of land cover and land use changes in China
China has experienced substantial land cover changes for hundreds of years, such as deforestation, agricultural expansion, and urbanization. These land cover changes can modify the physical and thermodynamic characteristics of the land surface, thereby influencing climate at regional or broader scales. In this study the Community Earth System Model (CESM) and Weather Research and Forecasting Model (WRF) were used to investigate the biogeophysical effects of land
cover changes over China. Both vegetation and urbanization were considered as land cover changes, and their impacts on climate in China were assessed separately and in combination. Results show that vegetation changes alone are able to alter the surface flux balance, therefore modifying regional temperature. Summer temperature changes are regulated by evapotranspiration and clouds, which can decrease daily maximum temperature but increase daily minimum temperature,
thereby decreasing the diurnal temperature range. Decreases in winter temperature over northern China are mainly influenced by surface albedo changes. Vegetation changes do not exhibit significant impacts on summer precipitation or the East Asian summer monsoon, but could strengthen East Asian winter monsoon and decrease winter precipitation in southern China. Compared to impacts of vegetation changes, the combination of vegetation and urbanization changes shows more
significant and extensive impact on temperature and precipitation in China. Urban heating produces greatly increased minimum temperatures and shows a remarkable warming effect in southeastern China, especially in winter when it offsets the cooling effects from vegetation changes. When urbanization is included, land cover changes can influence the East Asian summer monsoon and are associated with a significant southern increase and northern decrease pattern based on summer precipitation changes. Also, with the combined urbanization and vegetation changes, the East Asian winter monsoon is actually weakened, and exceeds the effects from vegetation changes. Therefore, we conclude that urbanization should be included in model simulations to provide realistic climatic impacts of land cover changes.
25 Feb: Jianjun Xu, George Mason University
3301 Exploratory Hall
Uncertainties in Estimation of Climate Changes
Global temperature trends coming from multiple datasets – including conventional observations, satellite retrieved products and climate model simulations – have been compared using the ensemble spread, which is a methodology of measuring uncertainty in climate changes. The results show that the magnitude of tropospheric warming and stratospheric cooling depends on the data sources, atmospheric heights, and geophysical latitudes. The spread in the simulations is much larger than that in the observations. The possible reasons will be discussed.
18 Feb: Edwin Schneider, George Mason University
3301 Exploratory Hall
The role of atmospheric noise in climate variability
4 Feb, 2:00pm: Danielle Wyman, George Mason University
163 Research Hall
Joint seminar with Center for Climate Change Communication
28 Jan: Ravi Shukla, George Mason University
256D/E Research Hall
Physical processes associated with the interannual dominant mode of regional and global Asian summer monsoon rainfall in NCEP CFSv2
This study examines the simulation of June – September mean (JJAS) of the Asian Summer monsoon in three different numerical experiments with a global climate model in which the atmosphere and ocean are coupled (coupled general circulation model, CFSv1), regionally coupled (“Pacemaker”), and with specified (SST) as AMIP/AGCM. Among the there simulations, we have found that amplitude and pattern of rainfall and wind at 850 hPa in the Pacemaker is superior to the other two simulations over the South China Sea, the subtropical western Pacific and the west coast of India. The observed relationships between NINO3.4 index and the Indian summer monsoon indices are remarkably better captured in the Pacemaker experiment than in the CGCM or AGCM/AMIP experiments. The northward/eastward propagation features and the spectral peaks (30-60 days) of rainfall are significantly more realistically captured by Pacemaker in comparison to CGCM and AGCM/AMIP.
Using observation-based analyses, we have identified that air-sea feedback processes associated with the leading interannual mode of the Indian summer monsoon rainfall originates within the northern Indian Ocean and is independent of the delayed and contemporary influences of ENSO. The improvements in representing the Asian summer monsoon features are evaluated using the latest CFSv2 during June to September.
Thursday 15 Jan, 11:00am: William Boos, Yale University
Research Hall, Room 256D/E
Monsoon depressions: dynamics of high-impact storms at the edge of the Tropics
Billions of people living in the monsoon climates of Asia, Australia, and Africa are highly vulnerable to floods and droughts. Much of the rainfall in these regions is produced by large (2000 km in diameter), propagating vortices that are neither typhoons nor classical extratropical storms, but another type of synoptic-scale disturbance that remains poorly understood. In June 2013, floods associated with one such vortex led to the deaths of about 6,000 people in northern India. This talk will provide an overview of monsoons and their constituent vortices, and will present recent results that are leading toward a new understanding of synoptic-scale monsoon vortices.
14 Jan: Tim DelSole, George Mason University
Research Hall Showcase
“The Roles of Aerosols and Natural Variability in Recent Climate Change”
The role of aerosols in recent climate change is controversial. Skeptics claim that climate models exaggerate the role of long-lived greenhouse gases. Some models suggest that what was previously thought to be natural variability may actually be due to aerosol forcing. This talk will review the debate about aerosol-forced climate changes and discuss new statistical techniques for separating aerosol-induced variability from variability due to long-lived greenhouse gases.
Monday 12 Jan, 1:00pm: Arindam Banerjee, U. Minnesota
3301 Exploratory Hall
Estimating High-Dimensional Statistical Dependencies: Advances and Potential Climate Applications
Estimating statistical dependencies between several variables of interest from a small number of samples is a key problem in statistical machine learning. The ability to find such dependencies can advance our scientific understanding of complex phenomenon in several domains. However, the confluence of `high dimensions,’ i.e., working with several possible variables which may be interacting with each other, and `small samples,’ i.e., a few examples of a phenomenon of interest involving the variables, makes finding such dependencies challenging.
In this talk, we give an overview of the dependency estimation problem, briefly discuss classical approaches based on linear correlation and mutual information, and present some recent advances on the problem. The key new insight is that as long as each variable interacts with only a few others, dependencies can be estimated from small samples. We illustrate the advances in the context of predictor screening for linear regression and inverse covariance matrix estimation for multivariate Gaussian copula models. We also discuss preliminary promising results of such approaches on climate multi-model ensembles. We end with some open
questions and challenges for dependency estimation in the context of climate.