Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
6 A Toolkit of Future Emissions ~ nventory Models IN THE PRECEDING chapters, the committee has taken a broad look at MOBILE. The report has discussed the uses of MOBILE in estimating mobile-source emissions, the technical issues associated with the model, issues associated with evaluating model uncertainties and accuracy, and alternative approaches for modeling mobile-source emissions. Two over- arching recommendations for improvements to the mobile-source emis- sions estimation process emerge from this review. The first is that the U.S. Environmental Protection Agency (EPA) should develop a long-term work plan, with input from the U.S. Department of Transportation (DOT) and others, on how to develop more accurate and effective mobile-source emissions modeling tools. The technical issues that should be addressed in this long-term plan are the focus of earlier chapters. The second recommendation, the focus of this chapter, is that EPA should develop a modeling "toolkit" that better serves the full range of cur- rent uses of the MOBILE software. The motivation for this recommenda- tion is that MOBILE is currently applied in situations for which it was not designed and is poorly suited. A "toolkit" of models is required, and here we lay out the structure for this proposed emissions modeling toolkit. The chapter proceeds with a discussion of the data needs and the user guid- ance that must accompany toolkit development. In closing, the chapter briefly touches on some of the institutional issues associated with the de- velopment and application of such a modeling toolkit. 797
7 98 MODELING MOBILE-SOURCE EMISSIONS REVIEW OF MOBILE'S USES AND SHORTCOMINGS MOBILE was originally designed to estimate mobile-source emissions inventories and compare the effects of control strategies. As dictated by various legislative initiatives, including the Clean Air Act, Intermodal Sur- face Transportation Efficiency Act, and the National Environmental Policy Act (NE PA), an increasing number of requirements have been placed on planning organizations to better assess mobile-source emissions and over- all air quality. As a result, MOBILE is now used for at least six general application areas (see Chapter 2~: 1. National and regional regulatory strategies. 2. Evaluation of control strategies and emissions inventories, and rate of progress. 3. State Implementation Plans' (SIPs) demonstration of attainment. 4. Transportation conformity and evaluation of transportation impacts in a nonattainment area. 5. Transportation control-measure effectiveness. 6. NE PA and evaluation of capital investments. These applications require assessment of mobile-source emissions at various temporal and spatial resolutions. Further, they often require in- teraction among the three different modeling disciplines: travel-demand, emissions, and air-quality modeling. Incremental improvements have been made to MOBILE over the years and MOBILE might still be well suited for aggregate regional and national analysis. However it cannot satisfy many of the applications listed above. Some of the individual is- sues associated with the use of MOBILE include the following: No protocol exists on the calibration of MOBILE model components. No protocol exists for standardizing emissions tests and ambient measurements made by the public and private sectors. No protocol exists regarding the evaluation of emissions-model esti- mates with air-quality measurements. No comprehensive assessment of MOBILE's sensitivity and sources of uncertainties has been completed that could help guide model improve- ments. . Federal regulations regarding SIP time horizons (i.e., short term) are inconsistent with air-quality conformity rules calling for air-quality assess- ments from transportation plans and programs 20 years in the future. MOBILE, used as a regulatory tool, does not project technology impacts on the time scale required for conformity analysis very well. With significant decreases in mobile-source emissions rates for typi- cal vehicles, off-cycle driving, high-emitting vehicles, and cold starts are
A TOOLKIT OF FUTURE EMISSIONS INVENTORY MODELS 7 99 important sources of urban emissions. Little testing data, however, are available to estimate these emissions. Travel-demand, emissions, and air-chemistry modeling data sets and results are collected for separate purposes, yet these data and results must be used in a common framework to assess the air-quality effects of mobile- . . source emissions. With on-road mobile-source emissions decreasing nationwide, the contributions of non-road emissions to urban inventories are increasing substantially. EPA is giving inadequate attention to estimating and vali- dating emissions from non-road sources. These deficiencies and the wide variety of applications of MOBILE, de- mand a fresh Took at setting priorities for work on mobile-source emissions modeling procedures. It is the opinion of this committee that MOBILE be supplemented with additional emissions modeling tools and data-collection efforts that win produce a better interface with transportation and air- quality models at the various levels of temporal and spatial resolution re- quired. DEVELOPMENT OF A MODELING TOOLKIT The proposed emissions modeling toolkit should have several compo- nents, as shown in Figure 6-1. These are an aggregated regional emissions-factor modeling component (i.e., the updated MOBILE moclel) for estimating emissions using aggregate vehicle activity data; a new-generation mesoscale emissions modeling component that inte- grates detailed transportation and emissions components to estimate re- gional and subregional (corridor) emissions and air quality through the coupling of vehicle operating conditions with appropriate link-based or trip-based emissions factors; a microscale instantaneous emissions modeling component that uses instantaneous operating conditions of individual vehicles to estimate con- tinuous vehicle emissions and can be used for a variety of applications, including generating emissions factors for microscale traff~c-simulation models, mesoscale emissions models, traffic data sets, and dispersion mod- els. If implemented, the committee believes these tools will provide the neces- sary broad suite of models that are needed for sound policy-making. They will enable better assessment of the health and environmental conse- quences of mobile-source emissions-control programs.
200 MODELING MOB`LE-SOURCE EMISSIONS Application' ~ Mobile-Source Emissions Modeling Toolkit - Aggregated Regional Emissions Model (MOBILE) Guidance Documentation (Protocol Manager) Integrated Transportation and Emissions Model Travel- New Generation Demand _ Emissions Model Model (trim or link-based) Microscale Instantaneous Emissions Model | · general purposes, stand-alone tool _ · integrated with traffic~ata and simulation models · integrated wig, dispersion models (Mesoscale) _ Regional Air-Quality Model l l ll FIGURE 6-1 Schematic diagram of the mobile-source emissions modeling toolkit. Regional Emissions-Factor Component MOBILES and its preceding versions have been developed as an aggre- gate estimation method that works reasonably weB for national and re- gional applications. MOBILE should remain as the aggregated regional emissions-factor component in the new suite of emissions models. This type of modeling component is required for comparisons of some control strategies and for comparing emissions from mobile sources with other source categories. It can estimate emissions inventories when provided with aggregated vehicle-activity data, as well as be used for evaluating new vehicle emissions standards, fuel specifications, and inspection and maintenance (~/M) program effectiveness. Further, it can be used to esti- mate the contribution of on-road vehicles to the nation's total pollutant emissions inventory and to monitor historical trends. This modeling com- ponent should be frequently upgraded. Data Needs for the Regional Emissions-Factor Component The data needs of the regional emissions-factor component of the model- ing suite should not differ drastically from that which is required in cur-
A TOOLKIT OF FUTURE EMISSIONS INVENTORY MODELS 20 7 rent MOBILE implementations. Typical user inputs include data such as vehicle technology, average speeds, ambient temperatures, fuel character- istics, tampering rates, fleet and vehicle miles traveled mix, mileage accu- mulation rates, registration distributions, basic exhaust-emissions rates, trip-length distributions, and operating-mode distributions. There are also assumptions about I/M program characteristics and credits and high emit- ters among others, as well as corrections for load, humidity, and air-condi- tioning effects. Metropolitan areas should be encouraged to develop their own input data for MOBILE, reducing the reliance on general default in- puts that might not represent local conditions. Mesoscale Transportation, Emissions, and Air-Qualily Modeling Component As discussed in Chapter 5, a new generation of mesoscaTe transporta- tion and emissions models is currently under development. These models offer much better promise for satisfying many of the needs required in the six application areas. With this component's ability to couple vehicle oper- ating conditions to traffic flow, a mesoscale transportation, emissions, and air-quality model has the potential to assess emissions impacts of a wider variety of controls and conditions than the less-detailed regional emis- sions-factor component. Calculating emissions inventories requires estimates of traffic flow and vehicle activity for different vehicle categories over the roadway network links. The mesoscale emissions modeling component would have a corre- sponding set of vehicle types and a set of emissions factors that correspond to different types of roadway facilities at different levels of congestion that can be applied on a link-by-link or trip-by-trip basis. These emissions fac- tors can be established through comprehensive testing and the application of the microscale instantaneous emissions modeling component, described in the next section. Through the simulation of fleet composition, conges- tion, time-of-day, and other parameters, the model wiB produce a spatially and temporally resolved emissions inventory. Pollutant concentrations for regional areas can be estimated for re- gional areas when the mesoscale model provides inputs to an air-quality model. This combination of models can then be used to demonstrate SIP attainment, such as for ozone, or used in determining whether major transportation plans conform to SIP requirements. An additional applica- tion at the mesoscaTe level is a detailed evaluation of I/M effectiveness. The mesoscale emissions modeling component should have sufficiently detailed temporal (hourly estimates) and spatial (1-5 km "ridded) alloca- tions that match well with current regional air-quality models. Because of the need for a high level of spatial detail, it is strongly recommended that
202 MODELING MOBILE-SOURCE EMISSIONS this mesoscale model component be developed within a Geographical Infor- mation System (GIS) environment. The mesoscale emissions modeling component should predict not just emissions inventories of nitrogen oxides (NO,j, volatile organic compounds (VOCs), and carbon monoxide (CO), but also fuel consumption, particulate matter (PM) and air taxies emissions, and greenhouse gas emissions (espe- ciaBy carbon dioxide (CON. As with the MOBILE model, the committee sees no need for separate models for different pollutants (e.g., a model such as PART separate from MOBILE to simulate PM emissions). Given an estimate of fuel consumption, the estimation of CO2 and some other greenhouse gases is straightforward. As described in Chapter 5, TRANSIMS is a major transportation, emis- sions, and air-quality modeling framework that has the potential to fill a portion of the requirements recommended for the mesoscale modeling com- ponent. TRANSIMS is an integrated system of travel-demand and emis- sions models simulating a detailed representation of a given population's travel behavior and the resulting emissions. However, the model win not be fully developed for several years. For this model, or any model, to be used in regulatory applications, it must undergo peer review such as the MOBILE model currently does. It also must be extensively validated and documented for users. TRANSIMS might fulfill this regulatory role; other new-generation models, such as the MEASURE and ITEM models de- scribed in Chapter 5, might also support this integrated transportation and emissions component in the toolkit. It cannot be overemphasized, though, that any model intended to fulfill such a regulatory role must un- dergo extensive peer review and validation, and provide in-depth docu- mentation to any potential users. Data bleeds for Mesoscale Transporlation and Emissions Modeling Component To estimate emissions inventories more accurately using the integrated mesoscale modeling component described above, will require a greater ar- ray of input data. These input data also require more spatial and tempo- ral resolution. This implies a need for more detailed data from state and local agencies. Some states and other users may find it difficult to develop the detailed activity-level data required to implement these models. How- ever, many agencies have already begun to build detailed local data sets, some of which are represented in a GIS framework. Transportation modeling requires better data in a number of areas (Chatterjee et al. 19971: Demographic data These data are important for determining trip-generation factors and include population by age, gender, and density;
A TOO[KIT OF FUTURE EMISSIONS INVENTORY MODELS 203 household size and number of dependents; and projections of these demo- graphic variables in future years. Economic data Important economic variables must be identified and developed, including number of households, income and vehicle owner- ship by household, employment, and projected future growth. Activity data These data help define the present and future num- ber, purpose, duration, mode, and other parameters for trips made by households. Land-use dataThese data are used for determining the effect of planning and zoning on transportation system utilization and perfor- mance. These data include residential and employment land-use fractions, concentrations of residential and employment land use, access from resi- dential and nonresidential areas to transit stops, and future land-use pro- jections. Roadway link data To effectively model the road network, data such as road segment length, number of lanes, posted speed, link capacity, and road-facility type, are needed. Transit data Both nonrail (i.e., buses and paratransit services) and rail transit data are required, such as round-trip travel times, average speed and stop times, stop locations, peak and off-peak service frequencies, and direction of service. Regional measures These data include production-attraction counts at zonal levels, VMT by road-facility class, mode splits, and prefer- ence surveys for transit and HOV lane usage. Microscopic measures Link level volume-to-capacity ratios, aver- age speed, travel times, transit ridership by hour, and percent of truck usage. Likewise, the emissions module wiB also require more detailed input data. It will require more spatial and temporal disaggregation of the same variables that are included in MOBILE including link-based emissions factors for different speeds and congestion levels, start distributions, and disaggregated rates for hot-soak, diurnal, resting-loss, and running-Ioss rates. This module will also require more detail on the vehicle fleet, such as registration distributions by vehicle class, fuel type, and emitter-level category. Microscale Transportation and Emissions Modeling Component A critical component to the emissions modeling toolkit is a microscale instantaneous emissions model. As described in Chapter 5, an instanta- neous or modal emissions model predicts emissions for a variety of differ- ent driving dynamics and can be used for a variety of applications.
204 MODELING MOBILE-SOURCE EMISSIONS . It can be combined directly with microscale traffic-simulation models (or measured data) to provide intersection and corridor-level emissions inventories. It can be used to reduce expensive dynamometer emissions testing. It can serve as the foundation for other components in the modeling toolkit. For example, given a set of roadway facility and congestion cycles, it can be used to determine link-based emissions factors; further, general- ized speed-correction factors (SCFs) used in the regional aggregate emissions-factor component can be improved with an accurate instanta- neous emissions model. Some of the applications that require an accurate coupling of transpor- tation and emissions modeling components can be performed using a com- bined microscale transportation and emissions model set. This is an area in which MOBILE has major problems. A microscale emissions model can be used to evaluate transportation control-measure effectiveness and some Congestion Mitigation Air Quality projects. Microscale transportation, emissions, and dispersion models also can be combined to determine pol- lutant concentrations for a variety of transportation projects. Data Needs for the Microscale Transportation and Emissions Modeling Component As described in Chapter 5, several instantaneous (modal) emissions models are being developed to predict second-by-second tailpipe emissions for a variety of vehicle and technology categories. The development of these models requires extensive vehicle emissions testing, with an empha- sis on real-world vehicle operation outside the performance envelope of the Federal Test Procedure (FTP). This detailed vehicle emissions testing should continue on a yearly basis to capture the effects of changes in fuel and automotive technology. Although instantaneous emissions models for light-duty vehicles (LDVs) are now becoming available, much more emis- sions data wiE be required to construct a comprehensive model for heavy- duty vehicles (HDVs). Another important component is determining the fraction of high-emitter in the on-road vehicle fleet as well as high-emitter emissions rates. Cooperative public-private partnerships in data collection are critical to expedite the development of necessary information. When the microscale emissions and transportation modeling component is combined with detailed traffic-simulation models, the models have simi- lar data needs as specified above, with additional data requirements such as signal types, intersignal spacings, signal phasing, mean start up delay at each signal, upstream distance of freeway exit signs from exits, and emergency-response vehicle times.
A TOOLKIT OF FUTURE EMISSIONS INVENTORY MODELS 205 Differences Among Emissions Modeling Components There are basic differences in the processes represented in the macro- scale, mesoscale, and microscale emissions modeling components. These differences affect how well each is suited for various applications. The macroscaTe emissions modeling component is the most aggregate in terms of coupling emissions with the operating modes and activity of vehi- cles. It uses parameters such as average vehicle speed, technology class and age, and VMT to estimate emissions for broad vehicle classes. The comprehensive nature of the model, representing emissions-causing pro- cesses for large numbers of vehicles, makes it useful for estimating large- scaTe emissions inventories (regional to national). However, the model does not directly link emissions to operating parameters that have a large impact on emissions, such as vehicle and traffic dynamics, and subsumes information about vehicle activity into parameters such as VMT and SCFs (see Chapter 3~. The mesoscale emissions modeling component contains a greater level of coupling between vehicle operating modes (i.e., vehicle activity) and emissions. The mesoscale level still uses "emission factors," but these are more closely coupled with travel-demand and traffic-simulation models to better represent the dynamic effects of traffic activity on emissions. Emis- sions can be associated with different kinds of driving, disaggregated for example by roadway facility type and by congestion level. There might need to be 50 to 100 cases of these mesoscaTe emissions factors that are indexed by transportation model links and different levels of activity. This is far more detailed than MOBILES, in which a single set of SCFs are ap- plied universally to all VMT, or MOBILES, in which SCFs will be applied to four facility types at varying congestion levels. The mesoscale compo- nent should provide a greater level of spatial and temporal disaggregation of emissions and incorporate a larger range of parameters that are known to affect emissions. The microscale emissions modeling component couples emissions with the instantaneous operating conditions of individual vehicles to produce a continuous (typically second-by-second) estimate of vehicle emissions. As such, it inherently handles emissions effects related to vehicle dynamics, and when coupled with microscale traff~c-simulation models, it predicts emissions related to traffic dynamics. The emissions from individual vehi- cles are summed to estimate total emissions for a particular traffic situa- tion during a particular time period. Because of the computational burden (and potentially large data-storage requirements), it is possible to do this only for relatively small scales, such as for intersections and corridors. Doing it for large regional areas is not feasible at this time. However, the instantaneous emissions modeling component can also be used to derive the mesoscale emissions factors. The microscale emissions model could
206 MODELING MOB!LE-SOURCE EMISSIONS essentially "precompute" the emissions factors by feeding in different rep- resentative driving cycles for the different links and congestion levels. Consistency Among Models in the Toolkit To allow for a nested progression of models, the models in the toolkit should be designed and maintained to be consistent, despite differences in the level of detail in inputs (e.g., variations in roadway network detail and meteorological variables) and the outputs (e.g., variations in emissions inputs over space and time). It is important that the different components in the proposed modeling toolkit are, to the best extent possible, based on the same data set and are able to predict similar emissions for similar con- ditions. A prominent example of how things should be consistent is that an instantaneous emissions model should generate the same integrated emissions numbers as an aggregate model for a specific driving cycle. Fur- ther, link-based or trip-based emissions factors can be generated directly from an instantaneous component, insuring consistency between those two layers. In general, the results of various emissions inventory methodolo- gies should be as consistent as possible. When they do differ significantly, the modeling application should be evaluated in detail to determine the reasons for the differences, and changes in the models should be consid- ered to reduce these differences. GUIDANCE DOCUMENTATION Guidance documents, suitable for the full range of expected users, are critically important and must be developed in concert with this modeling toolkit. These documents need to be specific about preferred methods and protocols in estimating emissions inventories for each of the six application areas. The guidance documents should not be developed solely by EPA, but rather it should include groups representing transportation, emissions, and air-quality disciplines. The MOBILE software has evolved over time in a rather haphazard way, seemingly backing into expanded uses in public-policy matters. The purpose of developing guidance or protocol man- agement in an interdisciplinary setting is to allow the development phase to be comprehensive, open, strategic, and responsive. It is a core require- ment for establishing trust in public-policy discussion as well as for ensur- ing broad and correct use of the models. Changes in modeling must be made in response to the demand users place on the current paradigm. The modeling toolkit approach presented in Figure 6-1 is specifically designed to expand to each of the applications. The guidance documents and established protocol management define how
A TOOLKIT OF FUTURE EMISSIONS INVENTORY MODELS 207 the emissions modeling toolkit should be applied for any specific applica- tion. Examples of the six general applications areas follow: Example 1: National and regional regulatory strategies. Work under this requirement would continue to be performed using MOBILE in the traditional manner. Example 2: Evaluation of control strategies, emissions inventories, and rate of progress. These evaluation requirements can be performed in a number of different ways, depending on the inherent level of detail need- ed. For example, in smaller communities that do not have overly sophisti- cated transportation systems, the traditional aggregated, regional emis- sions model component (i.e., MOBILE) coupled with a traditional travel- demand model will likely suffice. Communities that have complex trans- portation systems would apply the mesoscale transportation and emis- sions technique (see Figure 6-1~. Example 3: SIP demonstration of attainment. As in example 2, this work could be performed at different levels of detail. Example 4: Transportation conformity and evaluation of transportation impacts in a nonattainment area. Again, as in example 2, the same two approaches would be used to conduct these tasks. It is likely that large urban regions might opt for an integrated approach, whereas mid-sized urban areas might use stand-alone components of the modeling toolkit. Example 5: Transportation control-measure effectiveness. Most trans- portation control measure effects are seen primarily at the microscale level-of-detail. Therefore, the microscale components of the toolkit are to be used specifically for this purpose. The microscale components in this case would consist of a traffic simulation model (and/or off-model statisti- cal evaluations of modal activity changes) tightly coupled with the instan- taneous emissions model component. Example 6: NEPA and evaluation of capital investments. Again, the dual approach described in example 2 would be used; however, subarea values representing corridor conditions would have to be input into the traditional software. Thus, the modeling toolkit will add flexibility in responding to the six application areas. It also permits a choice of approaches to address the more extensive application areas, permitting some regions and urban ar- eas to use more integrated approaches, while others use a sequential, more aggregate approach to respond to air-quality requirements. SUMMARY OF POLICY AND INSTITUTIONAL ISSUES Many policy and institutional issues are associated with a toolkit ap-
208 MODE[/NG MOBILE-SOURCE EMISSIONS preach to emissions modeling. MOBILE plays a leading role in mobile- source regulation and, by extension, in air-quality regulation. The develop- ment of a modeling toolkit approach will require a major initiative cham- pioned by a broad constituency. The committee feels strongly that such an effort is warranted, given the significance of mobile-source regulation. It should be emphasized, however, that this requires a very broad and sub- stantial effort including coordination of planning, data collection, model development, documentation, and evaluation. EPA alone cannot develop an effective toolkit, and should not try to do so. There are many potential partners for EPA in this effort. Both the Cali- fornia Air Resources Board (CARB) and DOT have the modeling interest and expertise as well as the regulatory perspective. The committee recom- mends that within 1 year of this report's publication these entities initiate a cooperative program to develop a suite of models for assessing on-road and off-road mobile-source emissions. This program should begin with an assessment of the modeling needs that the regulatory and scientific com- munities foresee in the next decade. Discussions should include LDVs and HDVs, and should address all criteria pollutants, tonics, PM-2.5, and greenhouse gas emissions. Following the assessment of modeling needs, these agencies and other interested parties should assess the developmen- tal status of models that might be used within the toolkit. The product of this evaluation should be a plan for coordinating development of modeling elements. Finally, there should be an assessment of present and future data requirements. The emphasis should be on coordinating and standard- izing data-collection efforts and developing partnerships with universities, industry, and national laboratories to aid in this effort. The discussion of the need for a new suite of models, including a revised MOBILE model, raises the issue of creating a single regional emissions- factor mode! for use by aU states. California has its own legal authority to set automotive vehicle emissions standards and its own ambient air-qual- ity standards. Because of this, CARB has developed the EMFAC model, which is tailored to the automotive technology, fuels, and driving patterns in California (see Chapter 5~. As with MOBILE, EMFAC is updated peri- odically. The most recent version, EMFAC2000 shows substantially high- er estimates of emissions than the previous version of the model, EMFAC7G. With a gradual but perceptible closing of the gap between federal and California emissions standards, and hence automotive emis- sions-control technologies, the opportunity might arise in the future to fur- ther improve coordination of the EPA and CARB programs or even to com- bine MOBILE and EMFAC into one model. It is recommended that EPA and CARB immediately start to explore this possibility and to develop a time frame that is scientifically and technically appropriate. A joint report should be issued within a year of this report. Another critical institutional issue concerns model evaluation and vali-
A TOOLKIT OF FUTURE EMISSIONS INVENTORY MODELS 209 cation. The committee recommends that EPA and the other agencies in- volved in the development of a modeling toolkit not only undertake such studies, but also develop an institutional framework to ensure high-qual- ity, independent studies are conducted. The committee does not recom- mend any particular institutional arrangement, but several possibilities available. For example, in the past EPA has relied on blue ribbon panels and committees organized under the Federal Advisory Committee Act, such as the Mobile Source Technical Review Subcommittee discussed in Chapter 3, to provide advice on critical issues. Another example is the Emission Inventory Improvement Program (EIIP), which is a cooperative effort among state and local agencies, EPA, and industry. It was started by the State and Territorial Air Pollution Program Administrators and the Association of Local Air Pollution Control Officials to develop procedures for collecting, estimating, and reporting emissions data. A similar organi- zation could also be developed to coordinate model evaluation studies. A final issue involves the need for support for a modeling toolkit initia- tive from the legislative community that is involved in setting mobile- source regulations but not in implementing and evaluating these regula- tions. There is a constant demand *om this group for the quantification of emissions and air-quality impacts of mobile-source emissions-control pro- grams. Models are critical in this effort because they provide a consistent framework for evaluation and because they reduce the need to develop evaluation methods on a case-by-case basis. Quantifying the impacts of some initiatives, however, are beyond current modeling capabilities. This is especially true for programs that have small emissions impacts, little support data, or large collateral consequences. Too often, legislators de- mand assessments of controls with a high level of accuracy in situations where insufficient data and evaluations methodologies exist. And, too of- ten, regulators are unable to fulfill these requests. The difficulty of quan- tifying effects is often not appreciated by those who request evaluations. It is critical for legislators to understand the limitation of current model- ing capabilities and where additional resources are needed to further de- velopment of these capabilities. Improving this situation requires im- proved communications between those who perform evaluations of control programs and those who mandate such evaluations.
