General Background on the Weather and Climate Impacts Assessment…

General Background on the Weather and Climate Impacts Assessment
Initiative
Climate and weather create hazards and opportunities for society at multiple scales. However,
most of society does not have scientific expertise, and most scientists are unfamiliar with how
societal decision-making processes work. In the climate context, this process of bridging
between scientific knowledge and societal need is known as "assessment," while the weather
community might more familiarly call it "developing usable forecast information." Assessment
can be broadly defined as "the entire social process by which expert knowledge related to a
policy problem is organized, evaluated, integrated and presented in documents to inform policy
or decision-making" (GEA 1997). Assessments such as the U.S. National Assessment of the
Potential Consequences of Climate Variability and Change (USNA) and the international
Intergovernmental Panel on Climate Change (IPCC) assessment reports focus on synthesizing,
evaluating and reporting on what is known about climate variability and change and its impacts

Not all processes and products that fall under the rubric of assessment are the same however.
The NCAR Assessment Initiative focuses on impact assessments, a more narrow focus that aims
to assess the severity, likelihood, and effects of a given phenomenon, such as climate change and
extreme weather events, on a system of concern to society, such as agriculture, health or energy
supply. Within this area of research lie a number of critical scientific gaps that currently limit
our ability to effectively assess future impacts and provide quality information to decision-
makers. These difficulties include differing perceptions of uncertainty and extremes between
climate scientists, social scientists, and decision-makers; lack of tools for quantifying current and
future frequency of extremes; and so on (e.g., Moss and Schneider 2000, Webster et al. 2003,
Parson et al. 2003). This initiative concerns both filling these critical gaps and integrating the
different scientific disciplinary research necessary for informing decision makers regarding
current and future weather and climate hazards.

The Weather and Climate Impacts Assessment Initiative is organized around three themes:
characterizing uncertainty in all phases of impacts assessment, extreme weather and climate
events, and climate and health. It is mapped onto the following specific scientific objectives.

       ·   To quantify uncertainties related to multiple forcings (i.e., greenhouse gases plus
           land cover change, and natural forcings--solar variability, aerosols from volcanic
           eruptions) in climate models;

       ·   To characterize uncertainty on regional scales in climate projections that support
           decision-making;

       ·   To determine new robust measures of changes in weather and climate extreme events
           and their uncertainties (using extreme value theory), for extremes relevant to societal
           impacts;

       ·   To nurture an interdisciplinary research community to address the interactions
           between climate and human health; and
       ·   To work towards end-to-end integrated projects in extreme events and uncertainty
           that encompass physical science, impacts, and decision-making.

These objectives are fulfilled by a number of individual tasks (described below) that have been
selected because they address identified weaknesses of existing national and international
assessment processes such as lack of uncertainty estimates for climate projections, missing
elements of scenarios, or differences in the perceptions of the most appropriate way to consider
extremes (e.g., Moss and Schneider 2000, Webster et al. 2003, Parson et al. 2003).

NCAR is uniquely poised to study these topics, as it has a mission firmly grounded in the
atmospheric sciences, including climate and weather, as well as the responsibility as a national
center to provide science in service to society. NCAR also is staffed by renowned scientists in
these areas and has a multidisciplinary structure--the capability to mobilize scientists from
different disciplines around a central topic. This initiative is also of critical relevance for NCAR
because it provides timely, needed input to ongoing processes of national and international
importance--for example the IPCC, and future national and regional assessments.




                                                 2
     Renewal Proposal for the

   NCAR Weather and Climate

Impact Assessment Science Initiative




 Linda O. Mearns and Warren Washington




            September 6, 2002

 National Center for Atmospheric Research




                    3
Introduction
The WCIAS Initiative is built on three major themes:

   ·    Characterizing Uncertainty in Impact Assessment Science
   ·    Extreme Weather and Climate Events
   ·    Climate/Health Interactions

We organize this document around these three themes, which are described in full in the
Assessment Initiative Foundation Document (Mearns and Washington, June 2001). A number of
the projects are continuing tasks from FY01-02. These projects take priority over the new
projects. A separate attached sheet lists the priorities of all tasks in descending order. What can
be accomplished with level funding is indicated in the body of the document under the
continuing projects. Essentially, given the small amount of funding provided in FY01-02, very
little could be accomplished at that very low level. For several of the larger projects, efforts will
be made to seek complementary outside funding for some of their tasks.

I. Characterizing Uncertainty in Impact Assessment Science
The projects described in this Initiative element are highly diverse and cover many different
aspects of uncertainty analysis: uncertainty in climate model simulations with different emissions
forcings, uncertainty in future climate due to changes in land cover, exploration of the
uncertainty of past climates in climate models, incorporating uncertainty measures into climate
scenarios for impacts use, and uncertainty and decision making. Five projects covering these
topics are described below.

1. Uncertainty in Climate Model Simulations

Goals

The major goals of this project are to develop new techniques for quantifying uncertainty in
climate model projections and to apply theses techniques to recent transient runs of atmosphere-
ocean general circulation models (AOGCMs). Particular emphasis will be given to quantifying
regional uncertainty.

Progress to Date

In the first year of this project, transient Business as Usual (BAU) runs from the Parallel Climate
Model (PCM) were used to investigate aspects of the control run and climate change run in
reproducing some extreme events. Also, using a mixed model approach, progress was made in
elaborating on the uncertainty approach of Giorgi and Mearns (2002) in combining criteria of
validation and convergence in evaluating future climate changes using regionalized output from
9 different atmosphere-ocean general circulation models (AOGCMs) run with two different
emissions scenarios.




                                                  4
Research Plans

The research on uncertainty in climate model simulations will continue to be developed in three
main tasks.

The first one is aimed at the analysis of single model ensemble runs, using model output already
available to us (PCM all forcings runs for present day climate; BAU runs for end of the 21st
century climate; and Special Report on Emissions Scenarios [SRES] scenarios). Within this
context, the analysis of uncertainty will focus on issues of downscaling model output to the
observations' domain - and conversely "upscaling" station records to the model gridbox domain -
in order to correctly compare present-day climate simulations to available records and to sensibly
infer local impacts of simulated future changes. Aspects of climate change analyzed will be
indices of extreme events' intensity and frequency, influence of ENSO-like signals on them, their
trends and spatial patterns, and, for all of these quantities, a characterization of uncertainty under
precise statistical assumptions will be the central part of the analysis.

The second task will be the continuation of the analysis of multi-model ensembles following
Giorgi and Mearns (2002), with particular focus on determining regional measures of climate
change and their confidence levels. In this context, statistical models will be aimed at separating
within from between model variability through mixed effects models; assessing different factors'
relative importance (region/season/scenario/model and their interactions, for example) through
ANOVA-type analysis; accounting for outliers through robust statistics; investigating the
sensitivity of the summary measures of climate change to the scales of regional aggregation, by
use of spatial statistical models; and modeling the temporal evolution of processes through
multivariate time series analysis.

The third task concerns designing of experiments (model runs). A proof-of-concept will take
place using the idea of "daughters ensemble members." These will be short runs generated by
randomly selecting as initial conditions states at the end of ensemble members' long runs in order
to explore the "weather patterns" space under the same climate scenario. Underlying this
experiment is the more general issue of studying the relation between ensemble size and variance
of the estimates. A more complete approach will then be pursued, in which the theory of
experimental design will be applied, in particular ideas of latin hypercube-type design, in which
factors are varied optimally and efficiently in order to analyze their relative importance. We will
consider factors including changes to the emission scenarios.

(Note: Under level funding, only the second task could be accomplished).

Timeline of Accomplishments

YEAR 1
Results from tasks 1 and 2 will be available for specific study regions, together with a toolbox of
R software programs able to be easily adapted for similar analyses of different regions.




                                                  5
YEAR 2
Results of global analyses and full development of R software plug-in modules will be complete,
ready to be made available for the larger research community.

YEAR 3
Inter-model and inter-scenario comparisons will be performed. The results of task three will have
been analyzed with particular attention to the issue of modeling the propagation of uncertainty
stemming from emission scenario uncertainty.

NCAR Team
D. Nychka, C. Tebaldi, J. Meehl, L. Mearns, T. Wigley

External Collaborators
R. Smith, U. of N. Carolina, M. Berliner, Ohio State U., Chris Wikle, U. Missouri


2. Land Cover Forcing From the SRES Scenarios in Climate Models
Goal

To extend future climate change scenarios by including human impacts on land cover and soils.
This work has been initiated with LSM/PCM and will be expanded to use CLM/CCSM.
Simulated changes are land cover, human impacts on soil, and urbanization. These experiments
will address how human land use and land cover change are altering climate, water and carbon
cycles, and biogeochemistry. In particular it will address: (1) How have changes in land use and
land cover altered present-day climate, and how are they likely to alter future climates? (2) How
important is the land use and land cover change forcing relative to other IPCC SRES forcings
(e.g., greenhouse gases)? Parts of this work will be accomplished in collaboration with the
Biogeosciences Initiative.

Research Plan

A. LSM/PCM simulations

To assess the first-order impacts of land cover change, we propose a series of simulations using
PCM with the LSM land surface model. These simulations will complement existing climate
simulations of PCM using pre-industrial atmospheric forcings for 1870, transient historical
forcings from 1870 to present-day, and transient simulations to 2100 using the SRES A2
atmospheric forcing.

Proposed runs include: a pre-industrial atmospheric forcing equilibrium run to evaluate the
impact of present-day land use (existing control) against a natural vegetation land surface
representation; and transient simulations using the A2 SRES atmospheric forcings from 1980 to
2100 to evaluate the impacts of SRES-derived changes in land cover. By comparing runs using
natural vegetation, present land cover, and the A2 2100 land cover scenario (for the period 2065



                                                6
to 2100 only), we can evaluate the impacts of land cover change and the need for accurate land
cover information in assessing SRES scenarios.

B. CLM/CCSM development

While the LSM/PCM model provides a good first order estimate of land cover influence on
climate, there are a number of shortcomings to these simulations, especially the representation of
sub-grid scale land heterogeneity. We propose several new databases and model improvements
for the Community Land Model (CLM) used with CCSM to simulate different aspects of human
land cover change in more detail. These include:

   a. Land cover change

Goal: Create transient CLM land surface datasets to match those used in the IPCC SRES
scenarios.

Needs: Develop translation algorithms to convert SRES scenario land cover classes (biome
classification) to CLM-compatible plant functional types (PFTs), which allow for sub-grid land
cover.

Expected outcome: We expect significant impacts to both the hydrologic cycle and energy
balance on regional scales. We propose to develop a temperature fingerprint associated with
historic land cover change to compare to the observed surface temperature record.

   b. Agriculture

Goal: Include multiple crop types and implement aspects of the CERES crop models and the
CENTURY soil biogeochemistry and crop management to improve agriculture productivity
estimates and their impacts on climate. This will be in collaboration with the Biogeosciences
Initiative.

Needs: This work will proceed in several phases. (a) Develop additional agricultural classes
matching those used in the IPCC assessments. (b) Implement existing models of crop growth and
development (i.e., CERES crop models). (c) Develop and implement an irrigation
parameterization and a global irrigation dataset. (d) Develop and implement a full carbon cycle
for agroecosystems.

Expected outcome: Simulations of natural vegetation, and historical, present-day, and future
land-cover change will assess, in a consistent manner, the impact of agroecosystems on climate,
water resources, and the carbon cycle.

   c. Soil degradation

Goal: Assess the effect of human-induced erosion and soil compaction on regional climates.




                                                7
Needs: Use the GLASOD database and historical population trends to simulate past and future
soil degradation. CLM soil properties will be modified on the basis of these simulations.

Expected outcome: We propose to evaluate this impact on a global scale, and to identify a
temperature fingerprint.

   d. Urbanization

Goal: Evaluate the effects of urban heat islands on regional and global climate and hydrology.

Needs: Develop an urban canyon model as part of CLM. Develop an urban land cover database
linked to population density. Collaborative work in the Biogeosciences initiative will develop
anthropogenic biogeochemical emissions datasets in relation to population.

Expected outcome: Simulations will help understand the scale and nature of urban impacts on
regional and global climate. We will evaluate potential mitigation measures such as urban
planning policies that limit urban sprawl.

   e. Regional models

Goal: Much of land use and land cover change occurs at a fine spatial scale not explicitly
resolved by global models. This landscape heterogeneity is better resolved by regional climate
models, which are an important scientific tool for downscaling and impacts research.

Needs: Develop a common modeling framework and databases for global and regional climate
models.

Expected outcome: Reduction in redundancy and a common modeling framework for global and
regional models.




                                               8
C. Proposed modeling runs

In addition to the previously described LSM/PCM simulations, the following CLM/CCSM
simulations will address the climatic impacts of the new land cover specifications (bold script
indicates the difference from control):

Run     Type of       Objective                Sensitivity test                                   Status
        Run
CEC     22 years      Present day              Control Run                                        Exists
        Equilibrium   atmosphere
                      Present CLM land
                      cover
CE1     22 years      Present day atmosphere   a) Optimal number of urban classes                 Year 1
        Equilibrium   Urbanization             b) Sprawl vs no sprawl scenarios
CE2     22 years      Present day atmosphere   a) Impact of irrigation                            Year 1
        Equilibrium   Irrigation
CE3     22 years      Present day atmosphere   a) IPCC scenario comparison (e.g. A2 vs B2 etc.)   Year 2
        Equilibrium   Land Cover               b) Within scenario land cover uncertainty
CE4     22 years      Present day atmosphere   a) Optimal number of crop types                    Year 2
        Equilibrium   Agriculture              b) Prescribed vs interactive crop simulations
CE5     22 years      Present day atmosphere   a) Impact of degradation                           Year 2
        Equilibrium   Soil Degradation
CT1     1870-2100     IPCC Scenario            Evaluate multiple IPCC scenarios (e.g. A2 vs B2    Year 3
        Transient     Integrated land cover    etc.)

D. Convergence of impact models and earth system models ­ Workshop

With agriculture and urbanization as climate feedbacks, our model becomes more of an earth
systems model and overlaps greatly with the climate change impacts community, who are greatly
interested in the impacts of climate change on vegetation, agriculture, and urban climate. We
propose a workshop in the second year to address the convergence of impacts and earth system
models in the context of interactive vegetation (both natural ecosystems and agroecosystems)
and also possibly hydrology. We expect a workshop of 40 people, of whom about 10 would be
international.

Other Funding Opportunities

We will seek outside funding to partially support activities in the second and third years through
the NASA Land Cover Land Use Change Program.

Timeline of Accomplishments

YEAR 1
First-order sensitivity of IPCC simulations to land cover. Transient human population datasets.
Urban land cover parameterization and transient datasets. Irrigation sub-model and datasets.




                                                   9
YEAR 2
Soil degradation parameterization and transient datasets. Transient SRES land cover change.
Interactive crop parameterization using CERES crop models. Workshop on Impacts and Earth
System Model Convergence.

YEAR 3
Simulated climate with cities, soil degradation, and land cover change. Transient SRES
simulations. Merging crop model with Biogeosciences initiative.

NCAR Team
G. Bonan, L. Mearns, J. Meehl, K. Oleson

Internal Collaborator
Biogeosciences Initiative (Bonan on Project # 4)

External Collaborators
J. Feddema, U. of Kansas, R. Leemans, M. Schaeffer, RIVM, Netherlands


3. Climate Scenario Development and Distribution

Goal

To continue and expand upon NCAR's role as developer and provider of climate scenarios for
impacts research in the United States and internationally.

Progress to Date

The U.S. Workshop on Climate Projections, Uncertainty, and Climate Scenario Development for
Impacts Assessments was held in July 2002 at NCAR. An action plan on developing a Unified
U.S. Program on Scenario Development is being developed as the main output of the Workshop.

Research Plans

We will develop a data system to collect the most recent outputs from AOGCMS, minimally all
the climate model simulations from the SRES scenarios, develop appropriate baseline data sets
needed for combining with the climate model output, and incorporate measures of uncertainty
into the scenarios (perhaps even regional probabilities of the different scenarios). We will also
create a web-based tool that will allow for easy data acquisition of the scenarios, provide
guidance material on the use of scenarios, and provide links to other data distribution centers.
Particular attention will be given to providing scenarios based on projections from the three U.S.
Climate modeling Centers, NCAR, GFDL, and GISS. Also, collaboration with the Canadian
Climate Center is planned to include their modeling efforts. Regional climate model outputs for
domains covering all or parts of North America will also be collected and made available for
scenario use. The baseline climate data needed by impacts researchers will be developed in
collaboration with NOAA NCDC (T. Karl). Soils databases will be examined and produced on


                                                10
the same grid as the climate data. The soils databases of VEMAP will be the point of departure
for this activity.

(Note: under level funding, little could be accomplished on this project, beyond continued
development of the action plan and some minor work on gathering outputs.)

Timeline of Accomplishments

YEAR 1
Work will begin in mid-FY03. Collect outputs of existing AOGCM simulations using SRES
scenarios A2 and B2, from North American Climate Centers. Start Development of Web-tool.
Coordinate with NOAA NCDC the development of gridded climate database.

YEAR 2
Incorporate newer climate simulations from the other SRES scenarios, including North American
regional model runs, into the data distribution. Start developing guidance material for scenario
use. Complete development of Web-tool.

YEAR 3
Develop measures of uncertainty and incorporate into the scenarios. Provide guidance on the
measures of uncertainty and how they can be used in impacts assessment.

NCAR Team
L. Mearns, J. Meehl, D. Middleton, W. Washington, D. Nychka, T. Wigley

External Collaborators
R. Stouffer, GFDL. J. Hansen, GISS, T. Karl, NOAA NCDC, G. Boer, F. Zwiers, CCCma, R.
Street, E. Barrow, Environment Canada, L. Gates, B. Santer, LLNL


4. Climate Variability of Past Centuries ­ Regional and Climate Mode Response

Goal

This component of the strategic initiative is an extension to the CSENT project of the CGD-
Paleogroup, which focuses on natural climate variability during pre-industrial times. Previously,
Climate System Model (CSM) and Parallel Climate Model (PCM) simulations have been very
successful in reproducing climate variability and trends verified by the instrumental record, a
period strongly affected by the anthropogenic forcing. Extending the time frame several hundred
years prior to the appearance of this trend is crucial to verifying overall climate sensitivity.
CSENT performs experiments employing the fully coupled Community Climate System Model
(CCSM vers. 2.0) by sequentially adding important forcings including volcanic aerosol, solar
irradiance changes, land use and greenhouse gases. Because the period after A.D. 1600 is now
well covered by high-resolution proxy data of exceptional quality, comparing the best currently
available proxy records with coupled climate model results acts as an important test of the
model's ability to represent both past and future climatic changes. This is the first-order goal of


                                                11
this project. Three general focus areas are pursued.

Research Plan

a) Fingerprinting of Forcings

Results from the CSENT simulations will be compared to recent climate reconstructions on a
number of spatial and temporal scales. This work will be done as a data-model intercomparison
with Prof. M.E. Mann (U. of Virginia) and Prof. R.S. Bradley (U. of Massachusetts). We employ
new fingerprinting techniques, developed over the last year, which distinguish imprints of
climate forcings from internal natural climate variability. Results of such analyses using model
output with known specified forcings are compared to the relations from proxy climate networks.
Collaborators include Prof. P. Naveau (U. of Colorado), Prof. H.-S. Oh (U. of Edmonton,
Canada) and a PostDoc from the NCAR-GSP group led by Dr. D. Nychka. Short sensitivity
experiments with varying forcing series aim to increase the signal-to-noise ratio for periods such
as the Maunder Minimum and early 19th century.

b) Regional and Climate Mode Responses

North Atlantic Basin: The North Atlantic Oscillation (NAO), the dominant mode of winter
climate variability in the North Atlantic sector, is a natural mode well captured in atmospheric
GCMs. The NAO is one of the few large-scale modes that exhibit a clear dynamic response to
external forcing factors (solar, volcanic and greenhouse gases). Its low-frequency variability
component could suggest that the Little Ice Age episodes (LIA; ~A.D. 1300-1850) are an
expression of a positive NAO phase. However, there is debate within the paleoclimate
community as to whether there is indeed a clear fingerprint of the NAO present in land and
ocean proxies. Nevertheless, both sides claim a link to solar irradiance changes as a source of
variability in their reconstructions. Using CSENT experiments, we will investigate the role of the
NAO with and without external forcing, placing a particular emphasis on comparisons between
empirical paleo data and modeled patterns of forced temperature variation.

American Southwest (SW): Environmental conditions in the US-SW have always been strongly
controlled by climate, in particular, the availability of water. The richness of good high-
resolution proxy data (derived from tree rings) allows reconstruction of important environmental
parameters such as precipitation, summer temperature, drought, and fire occurrence. As a
regional application, we will compare the climate variability represented in the global coupled
experiments with the range observed in the historical and proxy climate record over the US-SW.
We particularly focus on drought conditions, which can directly be related to regional wild fire
danger. We will investigate controlling mechanisms for low-frequency changes responsible for
wet and severe drought conditions. To complement these investigations, a regional modeling
effort using the MM5/OSU nested into CCSM covering the conterminous U.S. and Northern
Mexico will be led by E. Small (U. of Colorado), building on his previous work to investigate
climatic patterns in higher spatial resolution. Forced with CSENT runs as boundary conditions
for a selected multi-decadal window, the regional model can better resolve local contributions
from, for example, topography and soil moisture, to the regional climate variability. Of particular




                                                12
interest will be the influence from land use and external forcing effects on the North American
Monsoon system.

c) Uncertainty Propagation

The overarching problem of uncertainty propagation in paleo applications (from the forcings to
the impacts) has not been seriously addressed in the literature. We will evaluate the robustness of
climate signals in both modeling and proxy data. On the one hand, we revisit fundamental
assumptions often applied in proxy reconstructions through comparison with globally available,
physically consistent climate model output. On the other hand, a number of "unusual" events of
past climates well captured by climate proxy information are simulated, using varying forcing
combinations and internal mode states with the GCM to evaluate uncertainty forcings and model.

d) Education and Outreach Component

We are collaborating and cost-sharing with the NCAR Education and Outreach Office (EO) to
develop K-12 educational activities on natural climate changes of the LIA that directly integrate
into the new Climate and Global Change exhibit (in preparation for FY03) for the Mesa
Laboratories. Inquiry-based exhibit components and worksheets for classroom use (K-12 levels),
as well as instructional support materials for teachers following National Science Education
Standards are developed, tested, and disseminated by EO and the PIs together with an education
consultant.

Requested Computer Time:

To produce the special runs required for this project (dedicated runs of the AOGCM for regional
model resting, regional model runs, etc.), a total of approximately 10,000 GAUs would be
required. While some of these GAUs can be accommodated through the CGD allocations,
approximately 5,000 GAUs will be needed from other sources. We are requesting approximately
5,000 GAUs from the Director's Reserve to cover primarily the regional modeling runs over the
three-year period. Most of the GAUs will be needed in the second year (see timeline below).

Additional funding opportunities:

We have identified a number of programs and special calls that we will pursue for
complementary funding of this project. These include: NOAA OGP, NSF-GEO-MATH for
aspects of the uncertainty analysis, and NASA RAs for the Earth System Enterprise Program.

Timeline of Accomplishments

YEAR 1
Completion of CCSM forcing runs and general analysis; temporal fingerprinting of Forcings,
estimate uncertainty; assemble database of North Atlantic reconstructions, characterize patterns
and their evolution; study of temporal ENSO/drought teleconnections in US-SW; drought-fire
link in SW from proxy data and control runs; Setup of MM5/OSU nested in CCSM.




                                                13
YEAR 2
Spatial fingerprinting of external forcing and internal modes (NAO); sensitivity experiments
using varying forcings; specified SST runs with North Atlantic basin cooling; 30-year runs of
MM5/OSU using control conditions; analysis of drought and fire in US-SW.

YEAR 3
Focus on uncertainty: test of robustness of global reconstruction techniques; investigate
mechanism for low frequency variability in North Atlantic (thermohaline circulation); 30-year
nested runs of MM5/OSU forced with landuse and external forcing; comparison of model and
proxy climate reconstructions for SW; depending on previous results, extension of analyses
using projection runs to AD 2100 with statistically prescribed external forcing; Winter:
workshop on integrated Little Ice Age research at intersection of models and proxy data.

NCAR Team
C. Ammann, H. Cullen, E. Wahl, D. Nychka, R. Johnson, S. Foster, L. Carbone

External Collaborators
E. Small (regional modeling): U. of Colorado; G. Bond, E. Cook and R. D'Arrigo (Lamont-
Doherty Earth Observatory, Columbia U.), H. Wanner (Swiss National Competence Center of
Research in Climate, Director), J. Luterbacher and C. Casty (Dept. of Geography, U. of Bern),
T.W. Swetnam (U. of Arizona, Laboratory of Tree-Ring Research), M.E. Mann (U. of Virginia),
and R.S. Bradley (U. of Massachusetts), P. Naveau (U. of Colorado), H.-S. Oh (U. of Edmonton,
Canada) .


5. Decision Making and Uncertainty: Managing Wildland Fire Risks: Climate
and Weather Information and Uncertainty

Goals

This program element will contribute to the development of a methodology for examining the
effects of uncertainty and the value of weather and climate information for the effective
management of wildland fire risks. The focus of the research will be on analyzing the roles of
uncertainty and information in situations where autonomous, but mutually interdependent,
decisions are being made by a number of individuals whose interests and objectives may
conflict. The goal of the proposed research is to contribute to the development of policy
alternatives, decision support tools and risk communication methods that could improve societal
management of these risks. This project entails major collaboration with the Wildland Fire
Initiative and some with the Water Cycle Initiative.

Progress to Date

Alison Cullen has spent time at NCAR assisting in the development of the Uncertainty and
Decision­ Making part of this initiative. Kathleen Miller attended a conference on decision-
making and uncertainty to gather information on the state of research in this area world-wide.


                                               14
Through these activities, Miller and Cullen developed the idea of starting the task on wildland
fires and decision-making.

Research Plan

 The incidence and significance of wildland fire risks as well as the costs and damages arising
during and after individual fire events are the result of decisions made by a large number of
public agencies and private individuals on many different time scales. Long-term decisions
regarding road construction, timber harvesting, vegetation management and investment in homes
and other built infrastructure affect the likelihood, intensity, and costliness of fire events. Near-
term fire-suppression decisions determine the net social costs and environmental impacts, as well
as the distribution of costs and benefits arising from current fire events. Post-fire land treatment
has further ecological and hydrological impacts, for example on aquatic ecosystems. Fire
suppression also affects future fire risks. (We will be working in collaboration with the Wildland
Fire and Water Cycle Initiatives regarding post-fire hydrologic impacts and land management
decisions.)

A systematic analysis of the interconnections among decision problems faced by all of these
many players would help to illuminate the nature of current controversies surrounding
implementation of the National Fire Plan, and could be useful in tailoring policies to best fit local
circumstances. Such an analysis also could help to identify the types of climate/weather and
other scientific information likely to be most valuable at various points in the decision process
and the most effective modes for transmitting that information to the appropriate decision-
makers.

Our proposed analytical approach will be to characterize decision environments as composed of
a set of interconnected decision trees. This will allow us to map out important nodes of
interaction among otherwise independent decision problems. This mapping technique can be
extended over any appropriate time-scale. Our working hypothesis is that conflicts, as well as
opportunities for productive negotiations and policy interventions, will tend to be clustered
around those points of intersection.

(Note, under level funding, this project will essentially not exist-- some continued project idea
development could be pursued through visits from A. Cullen).

Timeline of Accomplishments

YEAR 1
1) Survey existing work in this area. Identify linkages to other research efforts. Convene small
workshop to help map out research strategy and to develop a template or framework of elements
to be considered in looking at the role of climate/weather or other atmospheric science
information in the various decisions relevant to wildfire risks and impacts.

2) Create a mock-up model of the interconnected decision framework. Identify relevant software
for displaying and quantifying the decision problems, and examine the flexibility of the modeling
framework and the potential sensitivity of the overall social optimum to model specification



                                                 15
(e.g., number of independent actors, number and magnitude of spillover effects among their
decisions).

YEAR 2
3) Conduct case studies focused on two regions with contrasting characteristics (e.g., with
respect to: climatic fire regime/seasonality; level of development; susceptibility to post-fire
erosion damage; land ownership characteristics). This element will entail conducting on-site
interviews with relevant decision-makers in each area to identify variables affecting their
individual decisions, and their perceptions regarding the impacts of decisions made by others on
the costs and risks pertaining to their own decision problems.

4) Identify the critical policy issues facing each of these regions.

5) To the extent possible, quantify the range of uncertainty surrounding the important variables
in a set of policy-relevant decision problems for each of the two geographical areas.

YEAR 3:
5) Model the interconnected decision processes, using linked decision trees, in sufficient detail to
identify options for (a) improving coordination among decision makers and (b) optimizing the
flow of forecasts and other information to them.

NCAR Team
K. Miller, R. Katz, R. Wagoner (Fire Initiative)

Internal Collaborators
R. Rassmussen, D. Yates (Water Cycle Initiative)

External Collaborators
A. Cullen, U. of Washington; S. Kane, NOAA




                                                  16
II. Weather and Climate Extremes

 Research in climate and weather extremes is fundamentally motivated by their impacts on
society, which are considerable. The IPCC reports highlighted extreme events from both the
physical science and impacts point of view and strongly recommended more research in all
aspects of extremes. The initiative element on extremes consists of five projects, several of
which are continuing. The projects concern research in weather and climate modeling of
extreme events, downscaling of extreme weather phenomena, spatial scaling of extremes,
application of extreme value theory to atmospheric problems, and research on reducing societal
vulnerability to extremes. A particular effort is made through these different projects to integrate
across both atmospheric science aspects of extremes and societal concerns. Collaboration with
both the Wildfire and Water Cycle Initiatives is involved in several of the projects.

1. Extremes Toolkit
Goal

The goal is to develop software and a web-based tutorial for the fitting of meteorological
extremes in a form accessible to the broader atmospheric community.

Accomplishments to Date

In FY02, work was begun on the toolkit in RAP. Preliminary programming of appropriate
algorithms and preliminary development of a graphical user interface was completed.

Research Plan

A website will be developed that will make the software available and provide a tutorial on its
use, along with ample weather, climate, and impacts examples. The essential software in S-plus
or R will be developed, as well as the basics of the tutorial.

(Note: under level funding, this project would still be completed as described).

Expected Accomplishments

YEAR 1
Toolkit and guidance material and web interface will be completed (1-year project).

NCAR Team
R. Katz, D. Nychka

External Collaborators
R. Smith, U. of North Carolina, P. Naveau, U. of Colorado




                                                17
2. Extremes in Aviation-Related Weather
Goal

Extreme events, though of great societal importance, often are difficult to forecast using standard
objective forecasting approaches. Extreme value theory seems to provide a potentially useful
alternative approach for these prediction problems, which will be investigated in this study. In
particular, we propose to apply extreme value theory to four forecasting problems that are
currently under investigation in the Research Applications Program (RAP). Although research
and development in three of the four forecasting areas is supported through the Federal Aviation
Administration's Aviation Weather Research Program, this funding is for directed research, and
not intended for exploratory research such as that proposed here.

Progress to date

During the first year, the main focus of the extremes work was on initial development of the
extremes toolkit. In addition, in-flight icing was identified as the first area of focus for
applications of extreme value theory. Specifically, the problem of predicting or diagnosing icing
severity was selected as an area that could benefit greatly from this work. Initial datasets to be
used in this application were identified, and a general approach was developed.

Research Plan

The four research areas to be considered are (1) In-flight icing, (2) Turbulence, (3) Convection,
and (4) Public Weather forecasts. Each of the first three phenomena occurs quite infrequently
(e.g., over less than 5% of the continental U.S. at any one time) and can be directly thought of as
an extreme event. With respect to the fourth topic, it is difficult for statistical/objective forecasts
to correctly anticipate extreme values of variables such as temperature, winds, and precipitation.
Each of these areas has particular attributes that make them difficult to forecast, as well as
particular features of interest for forecasting. For example, one of the important aspects of icing
and turbulence events is the severity of the conditions; this aspect is especially difficult to
forecast due to the very limited observations of the phenomena that are available.

The four forecasting areas will be considered in the order presented above, through three stages:
(1) isolate the problem; (2) apply analysis techniques to appropriate data sets; and (3) incorporate
results into an improved forecasting algorithm. Stage one for each product will begin as the
previous topic enters the second stage. All stages will involve collaboration with a RAP scientist.
The extremes toolkit that is being developed as another aspect of this initiative will be applied to
the analysis of the data for these topics.

(Note, under level funding, only two of the four research areas would be considered, and the
project start would be delayed until FY04, after the completion of the extremes toolkit).




                                                  18
Expected Accomplishments

YEAR 1
Extreme value method for icing severity in an experimental version of the Integrated Icing.
Diagnosis/Forecast Algorithm. Methodology and data sets for analysis of turbulence data and
severity algorithm development.

YEAR 2
Extreme value method for turbulence severity in an experimental version of the Integrated
Turbulence Forecast Algorithm. Methodology and data sets determined for analysis of
convection data and forecast algorithm development. Analysis of convection data completed.

YEAR 3
Extreme value method for convection forecasting completed, insertion either into Autonowcaster
or other system. Methodology and data sets determined for analysis of public weather forecast
techniques. Analysis of forecasting data completed. Extreme value method for public weather
forecasts of extreme events in an experimental version of the Intelligent Forecast System.

NCAR Team
B. Brown, M. Politovich

External Collaborators
P. Naveau, U. of Colorado


3. Downscaling of Extreme Weather/Climate Phenomena

Goal

We will downscale severe thunderstorms (those containing large hail, strong wind gusts, or
tornadoes), using upper air variables from reanalysis, regional models, and global models. Using
the developed methodology, we will estimate changes in extreme climate phenomena under
various climate change conditions.

Research Plan

Severe thunderstorms pose significant threats to life and property around the world.
Understanding their geographic distribution and the potential for changes in that distribution are
important in helping people prepare for threats. The IPCC noted that we know little about the
future changes in these phenomena, which are not resolved in global or regional climate models.
Since they are rare events at any particular location, and the most extreme severe thunderstorm
events (e.g., violent tornadoes, 5 cm diameter hail, 120 km/hr wind gusts) occur infrequently
even in the locations where they are most common, the direct observational record of their
occurrence is of questionable value in estimating their true distribution. The best observational
record of severe weather events extends only back to the early 1990s for all severe weather and



                                                19
back to the mid-1970s for the most extreme events that, due to their impacts, tend to be observed
more reliably.

Over the past several years, databases of observed radiosonde ascents taken in the vicinity of
severe thunderstorms in the U.S. have been developed. These contain thousands of so-called
"proximity soundings," with approximately 1,000 associated with tornadoes rated F2 and higher
on the Fujita scale of damage (the 10% most damaging tornadoes), 2,000 associated with 120-km
and stronger winds, and 3,000 associated with 5-cm larger diameter hail. These soundings have
made it possible to develop relationships for environmental conditions (moisture, instability, and
shear) that are favorable for severe thunderstorms, and to discriminate between the environments
associated with different severe weather types. A recent study at the U. of Oklahoma, wherein
proximity soundings were derived, demonstrated that NCAR/NCEP reanalysis data are
sufficiently accurate to provide discrimination between weather types that is almost as good as
the observed data, but with more complete coverage than the radiosonde network provides. A
combination of four parameters identifies conditions that are more than 100 times as likely to be
associated with strong tornadoes than other conditions.

We propose to go beyond the observed and reanalysis project studies to look at the ability of
models to capture realistic environmental conditions associated with severe thunderstorms. In
the first stage, we will examine if soundings derived from the analysis cycle and forecasts out to
24 hours from a state-of-the-art mesoscale model, MM5, can discriminate between
environments. This will allow us to evaluate how well models that are well designed to handle
convective environments evolve in threatening environmental conditions. In the second stage,
we will extend the work to regional climate model simulations, in which the basic model
formulation is not tailored to convection. Finally, we plan to use recent results from AOGCMS
to look at the climatological distribution of severe convective environments in current conditions
and under climate change scenarios. The results of the previous studies, focusing on severe
convection in the U.S. will be applied on a worldwide basis, to see where and how often
favorable environments occur in other regions and how that frequency changes in climate change
scenarios.

Timeline of Accomplishments

YEAR 1
Develop and analyze 3-year climatology of proximity gridpoint soundings from 0-24 MM5
forecasts for continental U.S.

YEAR 2
Analyze 10-year archived regional climate model simulations from Iowa State U. for
climatology of favorable conditions for severe thunderstorms in and develop relationships
between model parameters and severe weather occurrence.




                                               20
YEAR 3
Analyze global climate simulations of current and future climate conditions in order to study
changes, if any, in frequency and geographical distribution of conditions favorable to severe
thunderstorms around the globe.

NCAR Team
J. Bresch

Internal Collaborators
R. Katz

External Collaborators
H. Brooks, NOAA/NSSL (Project Leader); C. Anderson, Iowa State U.


4. Extreme Events In Climate Models And Spatial Scaling
Goal

We will analyze the frequency and intensity of various extreme events particularly relevant to
climate impacts in regional and global climate models. Extreme value theory will be used in
these analyses. In addition, we will undertake a detailed robust examination of the spatial scaling
characteristics of extremes in climate models.

Research Plan

The statistics of extremes have only rarely been applied to evaluate the ability of numerical
models of climate to represent extreme events or their projections of changes in the frequency
and intensity of extremes. Comparisons of control run output with observations have been
hindered by a mismatch of scale between grid point averages and point measurements, an issue
that is particularly problematic for extremes and especially for precipitation.

We will work on relating extremes for point measurements to the corresponding ones for grid
cell averages, making use of bivariate extreme value theory. We will make a more systematic
study of the scaling properties of climate extremes, ultimately developing a fully spatio-temporal
statistical model. As a part of this activity simulations with the Convection Resolving Model
(CRM) as part of the Water Cycle Initiative (subtask 2e) will be used. Simulations will be
performed at a 3 km resolution for ten Julys for the U.S. Great Plains. The spatial scaling of the
extremes of precipitation from these runs will be examined and compared with those from
observations. It is assumed that the convective resolving simulations should scale similarly to
observations.

Comparisons of control run regional and global climate model extremes with observations will
be performed by the postdoctoral fellow, making use of the statistics of extremes and including
the scaling research to take into account the scale mismatch. Comparisons will also be made with
these coarser-resolution climate models and the CRM results decribed above. Changes in


                                                21
extremes simulated by a transient response climate model experiment would be estimated using
the statistics of extremes with time varying parameters. Specifically, climate model output being
used in the Water Cycle Initiative Task 1, Diagnostics Study of the Diurnal Cycle and
Precipitation, will also be used in our investigations.All these analyses would include explicit
statistical modeling of the annual cycle of extremes.

Timeline of Accomplishments

YEAR 1
Archiving and preliminary processing of extremes output from climate model control runs and
change experiments. Preliminary processing of extremes from daily precipitation and
temperature observations, as well as reanalysis data. In latter part of year, analyze output from
CRM simulations. Identify inadequacies in existing spatial scaling theory as applied to climate
extremes.

YEAR 2
Development of an application of bivariate extreme value theory to point and areal climate
extremes. Analysis of transient response of extremes in climate change experiments (regional
and global models). Development of spatio-temporal statistical model of extremes.

YEAR 3
Comparison of control run and observed climate extremes. Application of spatio-temporal
statistical model to climate extremes. Development of extremes toolkit module tailored to
analysis of extremes from climate models.

NCAR Team
R. Katz, D. Nychka, C. Tebaldi, M. Moncrieff (Water Cycle Inititiative)

Internal Collaborators
R. Wagoner, J. Coen (Fire Initiative), K. Trenberth, A. Dai (Water Cycle Inititiative)

External Collaborators
R. Smith, University of North Carolina, P. Naveau, University of Colorado


5. Influence Of Climate Variability And Uncertainty On Flood Hazard Planning
In Colorado
Goal

This project is intended to enhance the understanding and use of flood-related climate
information by the many groups involved in floodplain management along the Colorado Front
Range (CFR), including local officials and floodplain administrators, private consultants, state
and regional flood control authorities, and federal agencies. It has four research components: (1)
policy and decision-making, (2) weather and climate, (3) hydrology, and (4) impacts on flood
hazard planning; therefore, a multidisciplinary approach is needed.


                                                22
Progress to Date

In FY02, funding was provided to develop a detailed project plan and a proposal for submission
to NOAA/OGP Human Dimensions Program. The proposal has been submitted ,but results will
not be known until early winter.

Research Plan

(1) Policy and decision-making. The primary objective of this component is to understand the
decision-making processes in floodplain management at local, state, and federal levels; the
methods of using weather and climate information and dealing with uncertainty in that
information; and the problems perceived by users and stakeholders. An additional objective is to
aid providers of weather and climate information by explaining how new information is diffused
and new methods are adopted within the flood management structure. Key flood risk variables
used in flood hazard planning and mapping (such as discharge, depth, and velocity) will be
identified in discussions with floodplain managers and used in the other study components.

(2) Weather and climate diagnostics. Rainfall attributes over the CFR are affected by large-scale
climate forcings, especially during summer. The overall objective of this component is to
investigate whether estimates of flood risk variables can be improved through better
understanding of relationships between climate/weather regimes and summer precipitation on a
variety of temporal and spatial scales. Past flood events in the region will be associated with
meteorological scenarios that can be explored using hydrologic models. Meteorological
scenarios will also be related to larger-scale climate signals to evaluate the potential usefulness
of summer season precipitation forecasts for flood hazard planning.

(3) Hydrological modeling. The objective of this component is to develop a framework that
simulates plausible hydrologic scenarios consistent with the large-scale climate. A stochastic
weather generator, in combination with hydrologic models, will be used to generate realistic
scenarios of streamflow associated with the meteorological scenarios developed in (2). This
framework will be tested against data from past years and compared with flood conditions and
decisions made in those years. The resulting probability density functions will be used to
demonstrate how uncertainty in precipitation data propagates through hydrologic models,
increasing the uncertainty in estimates of flood risk variables.

(4) Impacts on flood hazard planning. Geographic Information System (GIS) technology will be
used to demonstrate potential effects of variability and uncertainty in climate information on
floodplain maps and risk assessments. Effective communication of the findings of this project to
stakeholders in flood hazard management and to providers of weather and climate information is
a primary objective. Results will be shared with stakeholders and researchers through
presentations and discussions at professional meetings, a website, and publications.

(Note: under level funding, the project probably will not be feasible. Reliance on the external
funding possibilities would be complete. Funding for M. Downton to further formulate research
plan would be maintained.)



                                                23
Timeline of Accomplishments

YEAR 1
Interview officials and technical experts in floodplain management. Prepare report on decision
processes and information use in flood hazard planning. Obtain hydromet data for case study
areas; develop meteorological scenarios; develop stochastic weather generator and generate
weather realizations. Calibrate hydrologic and hydraulic models. Develop database of land-use
and land surface characteristics.

YEAR 2
Evaluate impact of uncertainty by comparing simulations of flood risk variables associated with
different design rainfall estimates. Verify simulated streamflow against observational data; refine
meteorological scenarios. Enter weather realizations into hydrologic model to generate simulated
streamflow realizations. Incorporate socioeconomic data and hydrologic information.

YEAR 3
Discuss project results with flood management professionals; obtain feedback; present findings
at professional meetings. Develop scenarios to evaluate impacts of land-use change and large-
scale climate impacts. Test skill of model results through hindcasting; compare with actual
decision-making process. Develop demonstration of flood potential under varied conditions.of
model results through hindcasting; compare with actual decision-making processes. Combine
output of communications, diagnostics, and modeling to provide seasonal updates of flood
hazard in case study areas.

NCAR Team
M. Downton, H. Cullen, R. Morss, O. Wilhelmi

External Collaborators
B. Rajagopalan, U. of Colorado




                                                24
III. Climate/Health

1. Development of a Climate/Health Program

Goal

NCAR plans to develop a unique interdisciplinary research and education program via a health-
climate collaboratory. It will bring together leading institutions in health and climate science.

Research Plan

The area of human health impacts of climate is an extremely complex one, which requires the
interdisciplinary efforts of health professionals, climatologists, biologists, and social scientists to
successfully analyze the myriad relationships among physical, biological, ecological, and social
systems relevant to health impacts. This area is an endeavor where an integrated assessment
framework is obviously most needed. It is particularly important that as complete and rigorous
knowledge as possible regarding health/climate interactions be obtained in the context of all
appropriate technological and social considerations. The flurry of commentary over the past few
years on health/climate connections and the dangers of oversimplifying relationships, as well as
the recent NRC panel report that addresses issues in human health and climatic variability,
reflects this importance.

Much progress towards enhancing institutional coordination and collaboration to enrich
climate/health research has been and will continue to be brought about through development of
specific multidisciplinary projects. However, to effect more complete progress, an ongoing
multi-year program in health/climate issues would be particularly desirable, since problems in
appropriately developing, needed health/climate linkages often transcend the specifics of any one
study. In essence, specific studies in the context of a broader climate/health program may be the
most effective context for rapid development of this complex impacts area. The NRC Report
explicitly recommends the establishment of interdisciplinary climate/health programs to foster
research, training, and appropriate communication to policy-makers on issues concerning climate
and health.

Research Program Plan

NCAR plans to develop a unique interdisciplinary research program via a health-climate
collaboratory. It will bring together leading institutions in health and climate science. Such a
program is necessary to properly train individuals in performing the kind of complex
interdisciplinary research necessary to successfully tackle research questions in climate/
health/society interactions. Thus, the program will include a postdoctoral training program that
will produce the first generation of scholars dedicated to an integration of the health and climate
sciences. We envision a multi-year program that would be dedicated primarily to establishing
interactions between climate scientists and health scientists. Interaction among these groups,
social scientists, and ecologists is an additional goal. Three main institutions will initially form
the center of the Program: the National Center for Atmospheric Research, Johns Hopkins
University, and the Center for Disease Control and Prevention.


                                                  25
Program Elements

We will consider studying a wide range of issues in climate and health, ranging from heat-related
mortality and morbidity to complex models of vector-borne disease (e.g., malaria and dengue
fever). Consideration of interactions between climate variability (interannual fluctuations in
climate) as well as longer-term climate change and human health effects will be included in the
program plan. One focus of the program will include careful research regarding dangers of
extrapolating research results across different temporal scales. In conjunction with the research
interests within NCAR/ACD, health effects of increased UV-B radiation and air pollution will
also be considered as initial research topics. In addition to fostering in general greater interaction
between the two main communities, the following specific goals will be sought:
        1) Evaluation of vector-borne disease and other health-climate models, with particular
attention to their completeness in how climate effects are represented and their adequacy for use
in a climatic change context; this will also entail the determination of the completeness of the
models in representing the linkages between climate, disease incidence, and the complex societal
and environmental contexts of the interactions of all factors (societal, environmental, medical);
        2) Examination of and inclusion of the modeling of other resource systems (particularly
agriculture and water resources) in conjunction with the disease models;
        3) Examination of the adequacy of climatological databases for use in human health
models and development of recommendations for extending the climatological and
epidemiological data bases necessary for performing effective research in climate/health
interactions;
        4) Examination of adequacy of field data collection and measurement techniques for
model development and evaluation;
        5) Formal scholarly exchanges between NCAR and other institutions, such as Johns
Hopkins, and the Center for Disease Control (e.g., summer visits at NCAR);
        6) Establishment of a formal postdoctoral program at NCAR for health researchers and
climatologists interested in health problems;
        7) Visitor program for policy-makers and public health workers to become more
informed regarding climate and human health interactions.

Initial Development Plan and Timeline

The most essential starting point for this project is seeking out and hiring a climate-health expert
to direct the project. Through consultation with some of those experts interested in participating
in such a program, it was determined that having such an expert in place at NCAR was essential,
and that such a person should be supported on NCAR funds. Seeking out and hiring such a
person would be the main activity of FY03. Then, toward the end of FY03 or beginning of
FY04, a meeting of the interested collaborators would be held at NCAR and a detailed program
plan would be developed. The program plan would be submitted within the existing RFP cycles
of interested funding agencies or presented directly to interested program managers. It is thus
assumed that the program would be funded from outside funding but that the project leader
would be funded from the Assessment Initiative for the full three years. Other details of FY04
and 05 activities will be developed through the project leader and external collaborators.




                                                 26
NCAR Team
R. Harriss, M. Glantz, L. Mearns, S. Madronich

External Collaborators
The following individuals have already been contacted and have expressed and interest in
participating in such a program: Dr. Duane Gubler, Director, Division of Vector-Borne
Infectious Diseases, Center for Disease Control and Prevention; Dr. Dana Focks, USDA,
Gainesville, Florida; Dr. Joan Rose, Department of Marine Sciences, U. of S. Florida; Dr.
Jonathan Mayer, U. of Washington; Dr. Jonathan Patz, Johns Hopkins U.; Dr. Peter Webster,
Georgia Tech; Dr. Elizabeth Whitcolmbe, Royal Hospital for Neuro-disabilities, London, U.K.;
Dr. Mark Wilson, U. of Michigan, Michael Sharpe, Health Canada. The collaborators include
experts in epidemiology, disease-society interactions, microbiology and water-borne diseases,
environmental health sciences, and climatology.




                                              27
Attachment A

Project (and Budget) Priorities: FY03
The priorities are organized by project and then also by the different level of funding. The levels
include 1) Level funding, the same level as FY02; 2) Low, which is the a level generally higher
than the FY02 level, but one in which the projects have been reduced from their high levels to
something close to "bare bones"; and 3) High, which is the desired amount to fully complete the
projects as they are described.

The effects of level funding are presented in parentheses in the main narrative at the end of each research
plan. The research plans described in the narratives assume the high level of funding. The low-level
funding usually involves reducing some tasks in the various projects, but we do not go into those details
here. It is assumed that such details will be relevant in discussions of the Initiative.

The key final numbers to be considered are the sums of priorities 2-6, which essentially provide
the total needed to fund all projects at the low level except for the climate health program
($994,388). Including the health climate program raises the amount to $1,091.588. The amount
for funding all projects at the high level (for which detailed figures are provided in the budget) is
$1,544,786. Low-level amounts for FY04 and FY05 are also being calculated and are available
upon request. On average across budget re