Goal 1: Improve Understanding of the Atmosphere, Earth System, and Sun
NCAR Strategic Priority 2: Investigating the Interactions of the Atmosphere, the Broader Earth System, and Human Society
Over the past several years, the collection and analysis of data on aerosols, clouds, and storms has been an important element in the evaluation of rainfall enhancement programs that RAL conducts throughout the world. The primary objective of all these efforts is the better understanding of the natural variability of aerosols and clouds for the evaluation of the potential for rainfall enhancement in a region. Independently, the effects of aerosols as agents of significant climatic perturbations, particularly with respect to precipitation, have received increasing attention over the past decade or so. Understanding the potential for a direct or indirect aerosol effect, both processes involve changes in cloud microphysical processes, has become an additional factor to consider in documenting cloud and precipitation characteristics in regions proposed for weather modification activities. The feasibility studies on rainfall enhancement conducted by RAL examine both small-scale effects by direct weather modification and the large-scale impacts of indirect effects on the climate scale.
FY2008 Accomplishments:Figure 1: Map of Saudi Arabia showing the distribution of annual rainfall for a 50 year period (1950-2000) based on rain gauge observations. The circulations indicate the location radar observations winter study (November – May) in the central region and for the summer study in the southwestern region (June – September). Figure 2: Map of rainfall observed by satellite (TRMM 3B43 product) over Saudi Arabia for a ten year period of January 1998 – January 2008. The rainfall pattern matches well with the long-term precipitation climatology shown in Figure 1.
Using observations collected from two weather radar networks (see Figure 1 for the location of the seven radars in the two networks) and research aircraft, RAL scientists have gathered information that substantially aids in our understanding of the climatology of Saudi Arabia, as well as characterizing the variability of precipitating storms. The climatological distribution of annual rainfall observed by rain gauges across Saudi Arabia is shown in Figure 1 and 10-year climatology observed by satellite is shown in Figure 2. Dryness is the prevailing climatic characteristic of Saudi Arabia except in the Asir region (centered on Abha in Figure 1), which receives annual rainfall > 300 mm due to its unique geographical configuration and the local mountains, and a secondary peak located in northeast, which is associated with winter precipitation. Rainfall in most of Saudi Arabia is < 200 mm, highly irregular (i.e., large natural variability), and hence is not dependable. Analyses of the field data has continued into FY08. A second year of more extensive field studies of air chemistry, aerosol, cloud microphysics, precipitation, and storm characteristics was conducted for the period of December 2007 – May 2008 in the central region and June – September 2008 in the southwest (Asir) region of Saudi Arabia.
A rainfall enhancement assessment study has been conducted in Mali, West Africa for the past three wet seasons (2006-2008). Aerosol and cloud microphysical measurements collected during the field program combined with NASA satellite and NRL aerosol model forecasts show that even though a major source of aerosols is related to dust transport from the Sahara, the Saharan dust is mostly confined to the northern areas of Mali (even during monsoon conditions) with only an occasional penetration of dust further south. Figure 3 shows an example aerosol distribution for a week starting on 8 Aug 2006, which is a typical distribution observed during the rainy season. These studies also showed that local variations in aerosols embedded on the background levels may play an important role in the effects of aerosols on clouds. Storm trends have also been analyzed, improving our understanding of where, when, and how storms develop in this region. The diurnal distribution of the number of storms is shown in Figure 4 for the 2006 and 2007 wet seasons. Both years show a distinct afternoon peak. However, the 2007 season was much more active, which is likely related to the variability in large-scale dynamics and aerosols in the region. The 2008 field season focused on expanding the aerosol and microphysical observations for another season along with the initial implementation of a randomized cloud seeding study.
A new program to assess the feasibility of rainfall enhancement in Turkey was conducted in the region near Istanbul during the winter and spring 2008. Airborne data were collected during this period to study the distribution of aerosols, cloud physical properties, and the development of precipitation.
RAL has launched a new effort in Queensland this past year to scientifically investigate when and how well cloud seeding works to enhance rainfall from convective clouds near Brisbane, Australia. This was a unique field experiment because a network of advanced radars (dual-polarization and dual-wavelength) was implemented in the field in addition to well-instrumented research aircraft. In combination with the airborne measurements, the radar measurements make it possible to trace the physical chain of events from the natural or seeded small particles to droplet and ice crystal growth, subsequent precipitation development in clouds, and ultimately rain on the ground in both natural and seeding clouds.
A field campaign that utilized a C-Band polarimetric Doppler weather and an instrumented research aircraft was conducted in Eastern North Dakota for the period of 9 June – 11 July 2008, which coincides in the peak frequency of storms in the region. The objective of this field campaign is to better understand the effects of hygroscopic cloud seeding at cloud base on convective clouds in North Dakota. This project is a continuation of the original field program that was conducted in the summer of 2006, which indicated very positive results. Specifically, the project was conducted to determine if identifiable signatures of hygroscopic seeding in polarimetric observables or derived fields could be observed and to characterize hygroscopic seeding effects stratified by aerosol and CCN concentrations. The study was conducted randomly to understand cloud droplet distributions above cloud for seeded and non-seeded clouds to confirm inferences observed in the polarimetric fields.
The 2006-2008 studies on aerosols, clouds, precipitation systems, and the feasibility of cloud seeding in central and southwestern Saudi Arabia was the first step in developing the infrastructure and collecting data aimed at understanding the physical processes in rainfall generation in Saudi Arabia. The plan is to continue the studies in the same areas, but have more focused measurement campaigns to address specific objectives of the project. The development of a randomized seeding program is also planned for 2009. A training program for visiting Saudi scientists at NCAR is also planned for this coming year.
The next step for the 2008 Mali study is the analysis of the collected aircraft and radar data to better determine the natural aerosol and precipitation characteristics in Mali clouds, and the effect of cloud seeding on these processes and vice versa in the context of the randomized seeding cases. In addition, more training sessions for scientists and technicians in Mali are planned for FY09. An extensive randomized seeding program the will be conducted over the entire rainy season is being planned for June-September 2009.
RAL has launched a second field campaign for 2008-2009 to continue to investigate when and how well cloud seeding works to enhance rainfall from convective clouds near Brisbane, Australia. At the same time, analysis will be conducted using the data from the first field season to investigate the physical chain of events from the natural or seeded small particles to droplet and ice crystal growth, subsequent precipitation development in clouds, and ultimately rain on the ground in both natural and seeding clouds.
The focus of the project for FY09 will be on radar data analyses which include: radar data quality control, statistical analyses of observed seeded and non-seeded cases, and the generation of liquid water content, particle shape, and rainfall estimates will be examined and compared with the previous study. Also, the NCAR polarimetric hydrometeor identification algorithm will be applied to the observed cases. For each case (13 cases, in total), seeded and non-seeded storms will be categorized. A statistical database along with a physical characterization will be developed for each case. A case study analysis of aircraft data will be conducted to stratify the results based on aerosol, CCN, and/or cloud droplet observations.
A new program to assess the feasibility of rainfall enhancement in India will be conducted throughout the country of India starting in June 2009, which coincides with the start of the monsoon season. Airborne data were collected during this period to study the distribution of aerosols, cloud physical properties, and the development of precipitation. The project is expected to be conducted over a five year period.
The atmospheric boundary layer (ABL) is characterized by strong energy exchanges between the atmosphere and the land surface. As such, it can be sensitive to land-surface heterogeneity and cloud shading, which modulate that exchange. A better understanding of the ABL should lead to improved prediction of ABL structure and evolution, in turn improving information provided to weather-forecast users affected by ABL circulations.
FY2008 Accomplishments:Figure 1: Band-passed error kinetic energy at 10-m above ground, with bands centered, from left, on 80, 200, and 600 km. Red colors show a loss of predictability, and together the panels show the upscale transfer of energy responding to small-magnitude soil moisture uncertainty, with known spatial covariance, prescribed at scales from 8-64 km. Results are valid 24 h after perturbation, at 1200 LST (Central) 28 May 2002.
In collaboration with Penn State University, we used a hierarchy of models to study the response of the convective atmospheric boundary layer (CABL) to land-surface heterogeneity and clouds. A study was completed with the WRF model to quantify temporal and spatial scales of predictability of CABL winds as it relates to uncertainty in soil moisture uncertainty. The key findings were (1) at short time scales (tens of minutes) the CABL wind response is locked to the soil moisture uncertainty, but at longer time scales (hours) the phase of the response decouples; (2) the magnitude of error growth in CABL winds is primarily a function of the dynamics of the coupled land-atmosphere system, and much less a function of the uncertainty scale or magnitude; (3) the presence of deep convection can act as a mechanism for upscale transfer of errors in CABL winds, and can also increase the magnitude of the error; and (4) regardless of the presence of convective instability, nonlinearity in error growth is present at time scales from minutes to hours. An example of up-scale transfer of error from small to large scales, in response to soil moisture uncertainty, is shown in Figure 1. The presence of nonlinear error growth in CABL winds under a wide range of atmospheric conditions has obvious implications for transport and dispersion predictions. CABL wind-speed errors less than 1 ms-1 (below current mesoscale weather analysis errors) can easily double or more on time scales of minutes to hours. In terms of transport, these results imply the predicted location of a contaminant can be in error by a kilometer in less than an hour.
A single-column model using the WRF land-surface and ABL physics was used to begin studying the response of the column to stochastic clouds. Experiments were designed to quantify the effects of cloud stochasticity on winds and stability within the CABL. Using cloud observations from the Atmospheric Radiation Measurement (ARM) Central Facility at Lamont, OK, stochastic models for cloud base height and liquid water path were designed in collaboration with statisticians at North Carolina State University. The stochastic models were used to generate hundreds of time series of clouds, which were then applied as a boundary condition in the single-column model. We ran ensembles daily for the period May-July 2003, and analysis of those experiments is ongoing.
The EULAG LES model was also improved in preparation for studying the CABL within mesoscale gradients and subject to land-surface heterogeneity. The EULAG was coupled to the WRF model. Testing to improve the EULAG LES subgrid-scale statistics and resolved-scale profiles for stable nighttime conditions was completed. A new Monin-Obukhov scheme was introduced for calculating surface heat and momentum fluxes (Z. Sorbjan, personal communication) and was generalized for both stable and convective conditions, in the absence of topographic features. The EULAG LES then was verified against the weakly stable conditions of the Beare et al. (2004) inter-comparison project.
One key problem is the lack of turbulence at LES inflow boundaries when periodic boundary conditions are not used. An investigation of methods to address this began in FY08, and will continue in FY09. A hierarchy of successively more complex and realistic tests has been prepared to evaluate the selected approach, culminating in the heterogeneous lower boundary-condition problem.
In the current phase of experiments new EULAG LBCs have been created. The model uses specific inflow LBC conditions (e.g. mesoscale wind and temperature) with turbulence input in the form of random noise below the inversion layer. Uniform surface heating within the whole domain provides forcing for the generation of turbulence. In the standard version, EULAG first used specified outflow LBCs, which caused an undesirable flow reversal at the height of the inversion layer. The outflow conditions were changed to open LBC conditions with zero gradient normal velocity components, which successfully eliminated the reverse-flow problem. Additionally, sensitivity studies involving different surface heating, surface drag coefficient, ambient wind and changes in turbulence inflow conditions have been completed. Other options currently being evaluated include linearly extrapolated outflow LBCs and gravity wave absorbers for the reduction of gravity waves trapped in the inversion layer, and for improving the downstream velocity profile.
As preparation for real-data simulations to be conducted later, the LES must also be extended to support irregular topographical features across multiple scales. We began this work by applying the EULAG model with a periodic complex-terrain lower boundary. The model was set up to simulate up-valley and down-valley thermally driven circulations with idealized complex terrain during weak synoptic forcing. This work was included in a model inter-comparison project for idealized daytime flows as part of the Terrain-induced Rotor Experiment (T-REX) (Rampanelli et al. 2004). The objective of the model inter-comparison was to quantify the uncertainty of thermally forced flows in idealized and real valleys over a full diurnal cycle. Metrics of concern included net radiation, surface sensible heat flux, the intensity and evolution of the along-valley and cross-valley circulations (along-valley flux and vertical flux at the top of the valley, maximum velocities near the slope and valley center, etc.), and the evolution of surface temperature and winds.
Analysis of the single-column response to stochastic cloud base height and liquid water path will continue. Evaluations focus on computing ensemble sensitivities that are analogous to adjoint calculations. This effort seeks to quantify the sensitivity of CABL stability to both instantaneous and time-integrated cloud-base height and liquid-water path from the imposed cloud time series. It is expected that the sensitivities of the responses will be functions of the time over which the sensitivities are computed, and also of the integration length scale of the cloud properties. The time scale of these sensitivities will reveal the CABL’s memory of clouds, and which particular cloud parameters that can elicit a lasting response.
We plan to continue work to properly set up the EULAG LES to study the CABL with heterogeneous lower boundary conditions and mesoscale atmospheric gradients. Baseline simulations are being set up and run so that the method can be evaluated first against coupled mesoscale-LES simulations that ignore the explicit introduction of turbulence: (1) the initial baseline configuration will use WRF mesoscale initial and lateral boundary conditions, and homogeneous lower boundary conditions in the LES domain; (2) a second configuration will modify the surface fluxes in the LES in a simple way so that a spatially varying internal boundary layer develops.
To decrease the effects of transition from laminar to turbulent flow, establish fully developed turbulent flow as close as possible to the upstream boundary, and provide accurate downstream profiles, we propose implementing a variant of the perturbation recycling technique with rescaling. We will also consider a dynamic procedure to define the distance between the inflow boundary and the area with fully developed turbulence, with the goal of shortening that distance as the simulation progresses. Three problems that need to be addressed in this approach are the following: (1) dynamically define the downwind location in the plane from which the perturbations are copied (Liu and Pletcher 2006); (2) properly scale the perturbations to get the correct upwind turbulence inflow; (3) address the downscale TKE transfer from the mesoscale flow to the LES (Germano et al. 1991, Moeng et al. 2007). Work is ongoing to address these problems, and will continue in FY09.This figure represents sample results from Phase I of NARCCAP, wherein the regional models use boundary conditions from NCEP reanalyses. The top panel represents observed summer precipitation (CRU data set) and the lower panel displays the results of the simulation with the WRF model.
NARCCAP is the North American Regional Climate Change Assessment Program, an international program to generate high-resolution climate scenarios by nesting multiple regional climate models (RCMs) within multiple global coupled models (GCMs) over the coterminous United States, most of Canada, and northern Mexico. The resulting datasets will be highly valuable to climate change impacts assessment researchers because of their increased resolution and availability in GIS formats; and to investigators studying uncertainty in regional climate modeling by allowing the comparison of multiple models of the same future conditions. A large number of collaborators make up the project team: Within NCAR this includes scientists in IMAGe, CISL, CGD, and MMM. It also includes multiple university and national lab partners.
It is expected that over 40 TB of quality-checked data will be submitted for archiving. During FY08, work at NCAR has included: development of an automated testing suite to ensure that data is correctly formatted and uncorrupted; coordinating plans with colleagues in CISL to archive data into the NCAR Mass Store and publishing these data through the Earth System Grid; development of a Data Archiving Protocol document; website and end-user support; discussions and testing to ensure that end-users' needs are met; and collaboration with the modeling groups and quality assurance team at Iowa State University. The NCEP-driven runs (1979-2004) with all six RCMs have been completed and archiving of much of the data from these runs is completed. The completion of the GFDL time-slice experiment occurred last year, and near completion of the (Community Atmospheric Model version 3 (CAM3) time slice experiment has occurred in FY08. The first two sets of RCM runs driven by global coupled models (GCMs) have been completed and data processing is underway. See the figure for sample results from Phase I (NCEP-driven run).
The program climate simulations will be completed in FY09. At that time a total of 14 high-resolution (50 km) simulations over North America will be available to the climate change community.
The oceans have absorbed one-third to one-half of anthropogenic CO2 from the atmosphere, which is causing a significant decrease in seawater pH. Called “ocean acidification,” this process affects many ocean biogeochemical processes as well as many marine organisms and ecosystems from high-latitude planktonic systems to tropical coral reefs. Interest in ocean acidification has quickly become a major oceanographic research priority and has invoked strong political interest.
FY2008 AccomplishmentsVarious processes that need to be accounted for in modeling changes both the changes in seawater chemistry in coral reef waters, as well as the responses of reef organisms.
Work has continued at both the organizational level and in the field. NCAR interacts closely with NOAA researchers in leading a Subcommittee on Ocean Acidification within the Ocean Carbon and Biogeochemistry Program and recently completed a white paper calling for the formation of a U.S. National Research Program on Ocean Acidification, with specific recommendations toward its organizational structure. Much of this work builds on several U.S. congressional hearings highlighting ocean acidification as an important consequence of increasing atmospheric CO2 concentration. NCAR researchers also continue field work in Puerto Rico with NOAA and University of Miami, to establish a monitoring system for the carbonate system in seawater on the La Parquera coral reef. Work has also begun toward development of a numerical model that accounts for the interactions between seawater chemistry flowing over a coral reef, reef photosynthesis, and reef calcification.
FY 2009 Plans
A joint NOAA/RSMAS/USGS/NCAR field exercise will be conducted in May 2009, to test various components of a seawater chemistry monitoring design on La Parquera Reef, Puerto Rico. At a larger scale, a collaborative project with researchers at CNRS, France, is underway, which attempts to synthesize the contribution of shelf calcium carbonate production rates on continental shelves. This synthesis is proving to be a critical to understanding the global carbon cycle as well as providing baseline data for ocean acidification studies. The modeling of coral reef chemistry continues, with the intent of improving our ability to predict how ocean acidification will affect coral reefs at regional to local scales. NCAR will also continue to play a leadership role in shaping US research on ocean acidification.(a) Warming of average maximum SSTs between 1950 – 69 and 1987 – 2006; mean (large black dots) and standard deviations (vertical lines) are shown for 0.5 deviation at 0.1 locations (based on HadISST). (b) As in (a), but comparing 1950 – 69 and 1980 – 1999, based on one member of the 20th Century CCSM3 integrations (from Kleypas, Danabasoglu, and Lough, 2008).
Coral bleaching, a phenomenon by which corals expel their symbiotic algae, has affected nearly 40% of reefs worldwide over the last 25 years, and has caused mass mortality of corals in about one-third of these events. Large-scale coral bleaching events are strongly correlated with extremes in sea surface temperature (SST). This project addresses the spatial variability in extreme events and how they determine coral reef bleaching patterns.
FY 2008 Accomplishments
A study that included analysis of both observations and model output examined the reasons for the regional differences in coral bleaching in the tropics. This study pointed to the potential for negative feedback mechanisms to keep temperatures in check in the western Pacific warm pool, an area that also has had a low incidence of coral bleaching. While both observational and model data illustrate the slower rate of warming in this region in the 20th century, model runs for the 21st century show that the rates of warming will increase at about the same rate as other tropical regions. Another study examined whether differences in coral bleaching rates are related to differences in the natural variability of sea surface temperature that in turn affects a corals tolerance of temperature change. This study illustrates that corals living in waters with naturally low temperature variability tend to bleach at lower temperature thresholds than corals from more variable environments.
FY 2009 Plans
In a collaborative effort with the University of California, Berkeley and The Nature Conservancy, plans are underway to design a high-resolution modeling study of the Coral Triangle, the Australia-Indonesia region of high marine biodiversity. Pending funding, the goal is to use a Regional Ocean Modeling System adapted for the region to investigate the vulnerability of coral reefs to future changes in sea surface temperature. The factors determining susceptibility to bleaching will be assessed (e.g., exposure to temperature extremes, coral community composition), as well as the probability of reef recolonization by planktonic coral larvae following a coral mortality event. The goal of these work is to improve conservation efforts.