As Chapter 3 shows, there is considerable physical science knowledge and effort toward monitoring and forecasting weather in the urban environment—a product of decades of research. On the other hand, Chapter 2 is a more general, high-level, largely anecdotal summary of end users and their needs. The Committee believes that the uneven nature of the information in these two chapters is a testament to the richness of our knowledge about monitoring and forecasting weather versus the needs, understanding, perceptions, and uses of weather information by end users.
As noted earlier, information needs of users remain unmet, despite recent advances and emerging technologies that offer promise to improve urban meteorology prediction and monitoring. This chapter focuses on those unmet needs and on future directions in urban meteorology and monitoring. It examines short-term needs, which might be addressed with small investments but promise large, quick returns, and then explores future challenges that stand to require significant efforts and investments. Note that while short-term needs may be easily attainable, they are not just for short-term attention; instead, their full potential can be reached through long-term sustained efforts only.
Since urban areas are part of the land surface component in weather and climate models, key lessons from the history of land surface modeling, observations, and understanding can provide some guidance on the near-term opportunities and future challenges in urban meteorology.
• Land model details evolve with the treatment of other atmospheric processes (e.g., Manabe, 1969; Deardorff, 1978; Pielke, 1984; Dickinson et al., 1986; Sellers et al., 1986).
• Land model development benefits from model intercomparison (e.g., Henderson-Sellers et al., 1993).
• Land model development benefits from and motivates coordinated field experiments (e.g., Sellers et al., 1992).
• These developments help the establishment of a variety of important international research programs, such as the Global Energy and Water Cycle Experiment (GEWEX), the (earlier) Biospheric Aspects of the Hydrological Cycle (Kabat et al., 2004), and (the successor to GEWEX) the integrated Land Ecosystem-Atmosphere Process Study (iLEAPS).1
• These programs, in turn, substantially accelerate the progress in coordinated field experiments over different continents, covering all major ecosystems; land model development, including the explicit consideration of vegetation stomatal resistance and photosynthesis as well as the land atmosphere exchanges of energy, water, and trace gases (e.g., carbon, nitrogen); and improved understanding of land-atmosphere coupling (e.g., Koster et al., 2006; Zeng et al., 2010).
There are four clear short-term needs related to urban meteorology:
1. maximize observational data in different categories from diverse sources,
2. regularly update metadata of the urban observations using standardized urban protocols,
3. continue and expand international urban model intercomparisons over urban areas, and
4. develop and apply best practices to strengthen the dialog between urban meteorologists and end user communities.
The second and third short-term needs are related to research needs; the fourth one is related to the need for translating established science to practical applications for end users, while the first one is related to both research and translation needs.
One significant need in urban meteorology is for improved high spatial and temporal resolution observational data. To help address this need, short-term field experiment and long-term monitoring data would consist of
• forcing data for urban meteorological models (or land surface models in urban areas), such as near-surface air temperature, humidity, wind, precipitation, solar and longwave radiation;
• observational data to characterize urban areas and determine urban model parameters (e.g., roughness length, impermeable areas) and urban sources/sinks (e.g., anthropogenic heating);
• observational data for urban model validation; and
• long-term observational data for end users.
A variety of data in each category are available from different communities. The best source of forcing data in the urban boundary layer likely would come from previous field experiments in urban areas. The down-scaling from 32 km regional atmospheric reanalysis over North America (Mesinger et al., 2006) and coarser resolution (from 0.5 to 2.5 deg or about 50 to 250 km) global reanalysis (e.g., Decker et al., 2011) provides another possibility.
For the urban characterization and source/sink data, the necessary spatial coverage is provided by satellite and aircraft data, such as the Landsat 30 m land cover data, the MODIS (Moderate Resolution Imaging Spectrometer) suite of land surface data (e.g., land cover, vegetation index, surface skin temperature) at 250 m to 1 km resolution, high spatial resolution aircraft lidar digital elevation data, and survey and field experimental data for urban sources/sinks.
Model validation data are primarily from field experiments and surveys. Polar-orbiting and geostationary satellite data (e.g., surface skin temperature) are also crucial. The geostationary data can be available at 4 km and hourly resolutions. They are available up to about 60 degree of latitude; however, the spatial resolution decreases at higher latitudes.
Much work has been done in attempting to remove the urban effect in the long-term climate data record (such as the 2m air temperature) (e.g., Kalnay and Cai, 2003). To avoid urban contamination, only sites far away from current and expected future urban areas are selected by the U.S. Climate Reference Network. However, for urban studies, urban effects need to be included, and long-term urban data need to be developed from the Global Historical Climate Network (GHCN; Peterson and Vose, 1997).
Short-term need #1: maximum access to observational data in different categories from diverse sources, by
• securing access to existing data sets from previous urban campaigns (e.g., through central archives for existing urban data sets and corresponding metadata),
• assuring that long-term monitoring networks will serve needs of both the global and urban climate communities, and
• integrating data sets from various monitoring networks (e.g., from the National Oceanic and Atmospheric Administration [NOAA], the Environmental Protection Agency [EPA], the Department of Transportation [DOT], etc) into central data archives (e.g., at major cities) that can be easily accessed by the broader science and end user communities.
Compared with observations over natural vegetation, the representativeness of observational data over urban areas is a much more challenging issue. Observational data have maximum value only if they are accompanied by comprehensive metadata. Without detailed metadata, observational data over the heterogeneous urban areas could be easily misused by the urban meteorology community and others. Considering the heterogeneous nature of urban areas, the site selection (including measurement height) of individual instruments and comprehensive stations with multiple instruments needs to follow flexible guiding principles as discussed in the World Meteorological Organization (WMO) manual on urban observations (WMO, 2008). Quality assurance and management are also crucial. Additional examples of metadata are discussed in NRC (2009).
The network-of-networks study (NRC, 2009) has demonstrated that while a plethora of surface monitoring sites often exist in urban areas, metadata are typically lacking for these sites, data access is not easily available, and data quality may be questionable. Given the importance of this weather and climate data to end users, it is crucial to assess the suitability of these sites, and to collect and document metadata where appropriate.
Furthermore, because the urban environment evolves rapidly as development proceeds (e.g., in almost all major cities in China in the past 30 years, as well as in U.S. urban areas with rapid population growth, such as Las Vegas and Phoenix), metadata for these urban stations need to be regularly updated. For the same reason, rural stations would become urbanized, and their siting needs to be reassessed according to the guiding principle in WMO (2008).
Short-term need #2: regularly updated metadata of the urban observations using standardized urban protocols.
There have been numerous model intercomparison projects in the past three decades, such as the intercomparison of global atmosphere-land coupled models, regional atmosphere-land coupled models, land surface models, atmosphere-ocean-land coupled models, and paleoclimate models. In general, initial intercomparisons help in the discovery of major model deficiencies, such as the lack of energy and water balance in some land surface models, and identification of the importance of certain major processes, such as the role of vegetation in land modeling). These intercomparisons can be done using comprehensive observations or a combination of limited observational data with model output. They may also include sensitivity tests on hypothetical situations (such as the atmosphere-land coupled model sensitivity to a uniform increase or decrease of global sea surface temperature). Intercomparisons at a later stage usually provide a baseline for understanding model prediction and projection and associated uncertainties instead of focusing on model improvements. For instance, the land/atmosphere coupled global modeling project in Koster et al. (2006) documented the coupling strength of each model and identified the land/ atmosphere coupling hot spots.
Although the characterization of a natural vegetation type (e.g., evergreen needle leaf tree) does not change with horizontal scales beyond individual trees, urban characterization changes significantly with scale. For global weather and climate models with a typical grid spacing of ~30 km and ~100 km respectively, most of the urban areas in the world occupy a small fraction of a model grid cell, and only the gross features of urban areas can be considered (e.g., Oleson et al., 2010b). For continental and regional weather and climate models with typical grid sizes from ~2 km to ~30 km, more detailed urban models are needed. For local models over specific urban areas with grid sizes from tens to hundreds of meters, urban models need to consider different urban climate zones. For specific applications over buildings or a neighborhood (e.g., air pollutant dispersion) with a grid size of meters, urban models should take into account details of individual buildings.
Initial urban model intercomparisons have been done in recent years based on a dataset containing net solar and longwave radiation, sensible heat, and latent heat flux observations for an industrial area in Vancouver, Canada (Grimmond et al., 2010a). No model performed best or worst for all fluxes, however, some classes of models performed better for individual fluxes. Based on all statistical measures, the simpler models performed as
well as the more complex models, suggesting that we currently lack the physical understanding to develop complex urban environment models. The subsequent intercomparison involved four stages in which participants were given increasingly detailed information about an urban site for which urban fluxes were directly observed (Grimmond et al., 2011). Similar to the initial intercomparison, no individual model performed best for all fluxes. In general, a model will perform better when additional information about the surface is provided. Nevertheless, it is clear that poor choice of parameter values can cause a significant drop in performance for models that otherwise perform well (Grimmond et al., 2011).
The community is still in the stage of capacity building (i.e., developing the urban model for weather and climate models) and cannot yet systematically evaluate the urban model impacts. Therefore, it is not yet possible to present specific examples of deficiencies in current modeling abilities or to provide products that separate near-real-time forecasting from modeling for long-range planning. Operational weather forecasting centers have not separated forecasting evaluations over urban versus nonurban areas. This is a challenge for almost all countries. Even countries with urbanized weather and climate models do not consider urban versus nonurban areas in model evaluation and validation metrics.
Although initial intercomparisons have focused on the energy, water, and momentum fluxes that are needed for atmospheric models, possible future intercomparison could include trace gas fluxes (e.g., carbon dioxide) and aerosols (e.g., for air quality modeling, cloud microphysics modeling). Different urban areas (e.g., coastal, mountainous, tropical, etc.) could also be covered in future intercomparisons as well as urban areas in both developed and developing countries. Intercomparison efforts need to continue with comprehensive data or, when such data are not available, a combination of limited observations with model outputs.
Short-term need #3: continued and expanded international urban model intercomparisons over urban areas.
There are diverse end user groups with different needs in urban meteorology. There is a lack of communication within urban meteorology (e.g., between modelers and experimentalists), between different end user groups, and most importantly, between urban meteorologists and end users (Oke, 2006). One participant of the workshop (Appendix C), an experienced end
user in the department of transportation at a major city in the U.S., specifically mentioned that only half of the presentations by urban meteorology experts during the workshop were comprehensible. There are several reasons for this, such as the lack of communication experiences and the lack of understanding of each other’s needs, practices, and capabilities.
Although some end users only need to know the most likely outcome among different options (which is similar to a patient seeking advice from a doctor on a particular treatment), most users require probabilistic urban meteorological information that convey the uncertainty, or likelihood that an event will occur. There are uncertainties associated with any weather and climate prediction due to the uncertainties in initial conditions, boundary conditions, model parameters, model parameterizations of physical, chemical, and biological processes, and the chaotic nature of the atmosphere. Even with the same atmospheric conditions, different urban models may yield different results (e.g., due to different representations of vegetation and hydrology). Furthermore, given that numerous urban models do not perform well across all fluxes, they should be applied with caution, and users should be aware of the implications for decision making (Grimmond et al., 2011). There is also general difficulty in the weather and climate community (as well as the scientific community at large) in characterizing and communicating probabilistic information to end users (NRC, 2006). It is important to note that even though improved communication of probabilistic information is crucial, it is only one piece of the bigger picture. Working collaboratively with communication scientists can help bridge the gap between end users and urban meteorologists.
Short-term need #4: development and application of best practices to strengthen the dialog between urban meteorologists and end user communities.
Note that discussions here are mostly related to communication practices. Communication is also a social scientific field of study with many subfields. Subfields relevant to urban meteorology and end users include risk communication, science communication, organizational communication, and mass communication. There is an essential role of communication science (and other social science disciplines) in helping the meteorological community (urban or otherwise) understand the complex roles of people’s perceptions, attitudes, behaviors, and experiences in their using weather information for decision-making. For instance, communicating weather forecast uncertainty information is indeed important, but it is only one piece
of a much bigger picture of communication that involves how people process information, their past experiences, the channels through which they get information, the influence of what other people around them think and do, and so on. Fully integrating communication science (and other social science disciplines) in bridging the gap between urban meteorologists and end users requires significant efforts and investments.
Although there are some clear short-term and relatively straightforward steps that could be taken, other potentially valuable advances would require significant efforts and investments and would therefore likely be challenging to implement. The value of such activities would need to be weighed against costs, such as through socioeconomic analysis. For instance, to estimate the economic effects of weather variability in the United States, Lazo et al. (2011) defined and measured weather sensitivity as the variability in economic output that is attributable to weather variability, accounting for changes in technology and changes in levels of economic inputs (i.e., capital, labor, and energy). Eleven nongovernmental sectors of the U.S. economy were found to have statistically significant sensitivity to weather variability (as represented by temperature and precipitation), and the U.S. economic output is found to vary by up to $485 billion per year of 2008 gross domestic product, or about 3.4 percent, owing to weather variability. At the state level, the percentage can be above 10 percent (e.g., in California), as shown in Figure 4.1.
Socioeconomic research and capacity related to weather (including urban meteorology) were the focus of the earlier Board on Atmospheric Sciences and Climate (BASC) Summer Study Workshop in 2009 and the subsequent NRC (2010a) report, and hence are not included here. However, the point remains valid that through collaboration, the urban meteorology community and social scientists could develop a core interdisciplinary capacity for urban meteorology-society research and transitioning research to operations. They could start with three priority areas: “estimating the societal and economic value of weather information; understanding the interpretation and use of weather information; and applying this knowledge to improve communication, use, and value” (NRC, 2010a).
Through information gathering (including the workshop), deliberations, and expert judgment, the committee concludes there are three significant long-term challenges related to urban meteorology.
FIGURE 4.1 State sensitivity to weather variability as a percentage of total gross state product. SOURCE: Lazo et al., 2011. (c)American Meteorological Society. Reprinted with permission.
• How can new capabilities for urban observations be developed and implemented?
• How can weather and climate models be urbanized, and how can urban areas be included in the model prediction evaluation and validation metrics?
• How will the capability for integrated urban meteorology-decision support systems be developed?
New Capabilities for Urban Observations
Because it is challenging to obtain representative measurements from individual sites over urban areas, there is a strong need to develop new capabilities in two categories. First, technologies are needed to integrate the information that may be available from the network of personal digital
assistants (PDAs), including smartphones that are connected to the internet. Several available or emerging smartphone applications leverage input from the Global Positioning System (GPS) and cell phone towers to determine location-specific information, which could be useful for reporting and evaluating weather events. Next-generation apps may have the capability to deliver site- or condition-specific forecasts and nowcasts. Similarly, a network of vehicles with GPS capability (e.g., from the U.S. Post Office, United Parcel Service [UPS], Federal Express, and possibly taxi services) would be valuable in urban measurements (e.g., for temperature and pressure measurements). To have such networks, inexpensive sensors would need to be developed.
Secondly, given that urbanization affects the physical and dynamical structure of the planetary boundary layer (PBL)—roughly the lowest one kilometer of the atmosphere—new technologies for critical measurements within this layer are essential. The PBL influences both local weather and the concentration and residence time of pollutants in the atmosphere, which in turn impact air quality. Measurements in the PBL are also important for dispersion applications. The PBL is also the most understudied and undersampled layer in the urban atmosphere, in large part because of the difficulty of access over some parts of cities.
Existing and emerging measurement technologies are also discussed in the extended abstract of W.F. Dabberdt in Appendix A. One way to integrate data from various sources is through the urban reanalysis (Box 4.1).
Challenge #1: How can new capabilities for urban observations be developed and implemented, particularly using the network of PDAs (including smartphones), vehicles, and new technologies for measurements in the whole planetary boundary layer?
Urbanization of Weather and Climate Models
As the horizontal resolution in weather and climate models continues to increase due to advances in the models, urban areas can occupy a whole model grid cell or a large fraction of grid cell. The treatment of urban areas can be divided into three categories: (a) not considering urban areas by assuming the areas are covered by the dominant vegetation type in the grid cell; (b) considering urban area as a specific land cover type with specified parameters which can occupy a fraction of grid cell or a whole grid cell; and (c) urbanizing the land model in weather and climate models by considering more detailed urban processes (e.g., anthropogenic heat sources, impervious surface fraction). For the third category, urban model intercomparison has been done recently (Grimmond et al., 2010a, 2011). An urbanized land
model has been used in some weather and climate models, such as the UK Met Office Unified Model for weather prediction (Best, 2005) and the National Center for Atmospheric Research (NCAR) Community Earth System Model (CESM; Oleson et al., 2010b).
In the United States, urban areas as a land cover type are considered in the NCEP nonhydrostatic mesoscale model based on the Weather and Research Forecasts (WRF-NMM) over North America, with a horizontal grid spacing of 4 km at present. Urban areas are treated in more detail in the NCEP pollutant dispersion prediction. Urban areas, however, are not yet considered in the NCEP global forecast system (GFS) and climate forecast system (CFS). This may be related to the NWS policy of uniformity of service (i.e., weather service distributed without consideration of population density). In addition, because observing systems are more or less uniformly distributed, particularly where terrain is not an obstacle, the NWP evaluation and validation metrics do not consider urban areas explicitly.
The National Weather Service (NWS) has begun to shift to digital gridded forecasts. In particular, NWS just released the Weather-Ready Nation initiative, partly in response to 14 separate disasters (including hurricane Irene), each with an economic loss of $1 billion or more in 2011 (Figure 1.3).2
Challenge #2: How can weather and climate models be urbanized and how can urban areas be included in model prediction evaluation and validation metrics?
Only after addressing this challenge can the crucial question that is relevant to end users be addressed: how effective are these forecasting models over urban versus non-urban areas (including the spatial scales of such forecasting and associated error probability distribution)? Such models will also help address another question from end users: as cities grow, does the urban effect (e.g., on temperature, air pollution, precipitation) change continuously or abruptly?
Integrated Urban Meteorology and Decision Support Systems
It can be a challenge for scientists to understand one another, even scientists in different subfields of the same discipline (e.g., between modelers and experimentalists in atmospheric science). Therefore it is not surprising that there is a lack of mutual understanding between meteorologists and the
There are observational data from diverse sources for weather forecasting every day. These data are assimilated into the operational model to represent the four-dimensional state of the atmosphere (i.e., three-dimensional in space and one-dimensional in time) for the initial condition (i.e., the best estimate of the initial state of the atmosphere) in numerical weather prediction (NWP) and for model evaluations. These data are referred to as the “analysis” fields.
However, NWP models and the associated data assimilation system have been improved over time; thus the differences in the analysis fields from year to year can be attributed to the model upgrades. Furthermore, due to the computation-related time constraints associated with NWP, many data are not available for real-time data assimilation. Therefore, reanalysis—analyzing data using the same data assimilation system to integrate data from diverse sources—has been done in the past two decades. The first reanalysis came from the National Centers for Environmental Prediction (NCEP) and covers the period from 1948 to present (Kalnay et al., 1996).
Since then, other weather forecasting centers (European Centre for Medium-range Weather Forecasts [ECMWF] and Japanese Meteorological Agency [JMA]) as well as the NASA Global Modeling and Assimilation Office (GMAO) have released their reanalysis products. At the same time, NCEP has also released the regional reanalysis over North America with a higher horizontal resolution (32 km), which includes the direct assimilation of precipitation observations (see figure below). Furthermore, NCEP released the first atmosphere-ocean-land coupled reanalysis, in contrast to atmosphere-land coupled reanalyses that have resulted from other reanalysis projects. Today, reanalysis is widely regarded as a great success in atmospheric science (with the Kalnay et al. (1996) being the most cited paper in all geosciences).
While the 2m air temperature and humidity are assimilated during the warm seasons to adjust soil moisture in the ECMWF reanalysis, and surface precipitation is assimilated in the NCEP regional reanalysis, most of the land surface data (such as surface skin temperature, 2-m air temperature and humidity, 10m wind speed and direction, and precipitation) are not explicitly assimilated in various reanalysis projects. Therefore the land surface (or urban) reanalysis process involves
• adjusting the near-surface atmospheric fields (temperature, humidity, wind, precipitation, downward solar and longwave radiation) from reanalysis using surface observations;
• running the land surface models (or urban models) forced by these atmospheric fields; and
• obtaining land surface fields such as surface skin temperature, sensible and latent heat fluxes, upward solar and longwave radiation, soil temperature and moisture, and soil heat flux.
Prior land surface reanalysis efforts include the Global Soil Wetness Project (GSWP) (Dirmeyer et al., 1999) and North American and Global Land Data Assimilation System (NLDAS; Mitchell et al., 2004, and GLDAS; Rodell et al., 2004), both of which are produced by particular instances of the Land Information System (LIS) software framework for high-performance land-surface modeling and data assimilation (Kumar et al., 2006).
Following these successful efforts, it is important to develop local reanalysis over selected urban areas. Community efforts are still needed to address several relevant issues: What is the adequate spatial resolution based on end user needs and data availability? How can we downscale coarse-resolution data to obtain atmospheric forcing data and surface data to characterize urban areas? What urban land models should be used? What period should we focus on (e.g., the modern satellite era from 1979 to present)?
The NCEP Regional Reanalysis domain and its 32 km topography. Terrain elevation (m) is indicated by the color scale at the right. SOURCE: Mesinger et al., 2006. (c)American Meteorological Society. Reprinted with permission.
diverse end users found in urban areas. On one hand, meteorologists need to better understand how individuals and organizations interpret forecast information and integrate it with other inputs, such as socioeconomic, political, and cultural factors, in decision-making processes. On the other hand, end users need to better understand how observational data and models generate forecasts and associated uncertainties. Given that by the end of this century most people will be living in cities, this mutual understanding is crucial because of the two-way effects of urban development on local climate and local climate on urban development. This is even more challenging in a nonstationary world (Box 4.2).
Challenge in a Nonstationary World
With the global warming of the past century and the expected warming in the coming decades, the assumption of stationarity in weather and climate is widely recognized to be incorrect (Milly et al., 2008). This nonstationarity of events (e.g., fewer cold waves, more heat waves; drought periods followed by extreme precipitation and flooding) presents a major challenge for planning and engineering activities that have a time scale of decades or longer.
Improving modeling capability in a nonstationary world requires scientists to further improve modeling and observations of natural and anthropogenic processes over urban areas. Nonstationarity, however, also implies that good modeling skills in the past and at present do not guarantee those skills for the future. In other words, prediction uncertainties will be larger.
Therefore, there is a strong need for decision makers to learn to manage uncertainties associated with climate projections and projections of human activities. One way to do this is to use climate model sensitivity tests to identify potential vulnerabilities of proposed adaptation strategies. This would allow decision makers to systematically examine the performance of their adaptation strategies over a wide range of possible scenarios which are driven by uncertainties about future climate and several other economic, political, and cultural factors (Pielke, 2009). This urban meteorology-decision support process can then be iterated to develop a strategy that is sufficiently robust across various alterative future scenarios. A similar iterative risk management framework has been emphasized in NRC (2011).
For instance, today’s 100-year flood zone in a coastal city (such as New York City) based on historical data cannot be used for urban planning and climate change adaptation in the future. With rising sea level in the next few decades due to global warming, a 100-year flood would inundate a far greater area of this city, and the frequency of flooding and the impacts of storm surge would also be significantly increased. Therefore, urban planning and climate change adaptation (e.g., building storm-surge barriers) need to operate knowing that nonstationarity is the new norm, and planning should involve iterative interaction between scientists (working on climate projections) and urban and regional planners.
To meet this challenge of mutual understanding between users and meteorologists, a combination of complementary approaches is required. These approaches include urban testbeds, applied science projects that involve meteorologists and end users, and joint urban meteorology-decision support exercises (e.g., emergency response, climate change-urban planning).
The idea of an urban meteorology testbed was proposed in Dabberdt et al. (2005) and NRC (2010a): “a testbed is a working relationship in a quasi-operational framework among measurement specialists, forecasters, researchers, private-sector, and government agencies aimed at solving operational and practical regional problems with a strong connection to the end users.”
Testbeds typically result in more effective observing systems, better use of data in forecasts, improved services and products, and economic/public safety benefits, and in the end, more effective decision making by users. The translation of R&D findings into better operations, services, and decision-making is accelerated by testbeds.
There are ongoing efforts in a number of cities such as New York City (Reynolds et al., 2004; Arend, 2010; see Chapter 2) and Oklahoma City (Basara et al., 2010) to establish urban observation networks that provide long-term observations (see Box 4.3 for more examples). Combining a suite of in situ and remote sensing instruments to provide a detailed picture of the urban atmosphere in all three dimensions, as well as integrating observations and modeling in an urban testbed with strong stakeholder involvement, remain challenges that have not yet been met in any U.S. city.
The establishment of several urban testbeds across the country and worldwide was commended by several workshop participants as a move in the right direction. However, there is still a lot of work to be done. In addition to identifying additional observations needed for ideal coverage of the urban PBL and developing strategies for simultaneously advancing modeling tools, additional end user communities (e.g., on the business and urban planning side) need to be brought on board.
Discussions at the workshop further highlighted that criteria and methodologies for the design of three-dimensional urban networks are still not well established. The World Meteorological Organization guidelines for urban surface monitoring sites (WMO, 2008) are an important first step toward urban network design. In addition, studies are needed to identify crucially important data as well as ideal sensor deployment, which would not only facilitate long-term urban records but also improved urban forecasts. Observing System Simulation Experiments (OSSE), which have been successfully tested for radar networks (Xue et al., 2006), should be developed for urban monitoring networks.
One good example of an urban testbed is the Helsinki testbed in Finland (Koskinen et al., 2011). It has been established and maintained since 2005 by the Finnish Meteorological Institute and Vaisala. It is an open research and quasi-operational program designed to advance observing systems and strategies, understanding of mesoscale weather phenomena, urban and regional modeling, and applications in a high-latitude coastal environment (see figure below).
Another good example is Shanghai, which is also actively pursuing development of an urban testbed that includes developing advanced observed systems for advanced models and providing actionable information that addresses user needs. Shanghai also has a multi-hazard early warning system; and information is shared between various agencies.
The Center for Collaborative Adaptive Sensing of the Atmosphere (CASA) and the North Central Texas Council of Governments (NCTCOG) are in the process of establishing the Dallas-Fort Worth (DFW) Urban Demonstration Network (Appendix B). Users will include NWS and River Forecasting Center (RFC) forecasters and emergency managers, while users from other sectors such as transportation, utilities, regional airports, arenas, and the media will be added later. This effort will also develop the business models for federal/municipal/private partnerships that will sustain the network.
Observing sites in the domain of the Helsinki Testbed (right), and larger area model domain with nearby RAOB sounding stations (left). SOURCE: The Finnish Meteorological Institute, http://testbed.fmi.fi/Stations.en.html.
Finding funding and infrastructure mechanisms to establish and maintain urban testbeds will be challenging, especially given the need to integrate existing and new monitoring networks with modeling techniques and provide long-term urban climate data records and validated modeling output tuned for various end users. Databases of urban building structures and transportation networks should also be integrated and continuously updated.
Challenge #3: How will the capability for integrated urban meteorology-decision support systems be developed through the integration of
• support for future intensive urban research projects that integrate modeling and observations and focus on improving the fundamental knowledge of physics and dynamics in the urban atmosphere,
• increased dialogue between urban meteorologists and end users, and
• urban meteorology testbeds?
The field of urban meteorology has grown considerably in the past 50 years, and with the increased growth of cities worldwide, including the United States, there is a pressing need for continued scientific advances within the field. As the capabilities within urban meteorology have improved, the uses for urban weather information and its value to decision makers have increased. Users of urban meteorology information need it to be available in a wide variety of formats, within time constraints set by users’ decision processes. In order to help meteorologists provide this tailored information, there is a need for more direct interaction with key end user communities who can help identify their information needs. By advancing the science and technology related to urban meteorology with input from key end user communities, meteorologists will be better able to meet the needs of diverse end users.