Information Sources to Support ADB Climate Risk Assessments and Management:

This technical note provides information that supports climate risk assessment experts undertaking early stages of project development in Asia and the Pacific region. The Asia and Pacific region is vulnerable to extreme temperatures, flooding by heavy rainfall, sea level rise, coastal erosion, and damage by tropical cyclones. This technical note provides information that supports climate risk assessment experts undertaking early stages of project development in the region. The information is grouped into four major categories: inventories of national emissions, climate risks, vulnerability, and impacts; historic weather, climate, and environmental change; regional climate change projections; and climate change impacts and adaptation. The note also identifies opportunities for capacity development in key skills such as geospatial analysis, data testing and post-processing, regional climate downscaling, and impact assessment.


Executive Summary
This technical note is intended to support climate risk assessment (CRA) experts, in particular, those undertaking the early stages of project development. Time and resources could be saved by attaching this document to terms of reference issued to CRA consultants. However, there is a limit to which globally accessible, open source data can meet the detailed information needs of local adaptation projects. This note supplements rather than replaces efforts to gather relevant climate information from government agencies and counterparts, especially during the project concept phase.
About 70 sources of public information have been compiled for the Asia and Pacific region, including data on historical and future climate, climate-related disasters, indicators of national vulnerability, and preparedness to adapt. Additional resources produced by ADB are also incorporated. This is a living document-appendixes should be periodically refreshed as new information is located and quality assured. Data sources are collated in four appendixes that broadly map to successive phases of the ADB Climate Risk Management Framework. The appendixes cover: (i) national emissions, climate vulnerability, risks, and impacts; (ii) historic weather, climate, and environmental change; (iii) multidecadal, regional climate change projections; and (iv) climate change impacts and adaptation.
Most of the data identified are contextual-they provide high-level information about present and future climate risks at national and/or sector levels. Additional capacity development may be required in specialist "gatekeeping" skills, such as geographic information systems, data homogeneity testing, post-processing research data formats (e.g., NetCDF), regional climate downscaling, and impact assessment. By growing these technical capabilities, more data could be accessed from the same public platforms or combined in ways that add value to the CRA.
As well as strengthening the technical capacities of local consultants and CRA experts, access to existing research-grade data stores should be improved through closer cooperation with scientific programs. For example, new interfaces could be developed to open research archives to a broader user base. In particular, there is an urgent need to widen access to the daily and sub-daily climate information needed for economic analysis, engineering design standards, stress-testing adaptation options, and other aspects of project design.

Introduction
The Asia and Pacific region is vulnerable to temperature extremes, and flooding by heavy rainfall, sea level rise, and tropical cyclones (Asian Development Bank [ADB] 2017). Risks from climate change are further amplified by the pace of population growth and rapid development of urban areas. Both trends are concentrating people and assets in places that are potentially exposed to climate threats. Climate change also has implications for food, water, and energy security; migration; and the stability of trade networks in the region.
To counter these risks, ADB has made a commitment to provide $6 billion of climate financing per year by 2020, of which $2 billion will be allocated for adaptation.
Due diligence requires that adaptation finance is invested in projects that deliver intended benefits at the least cost. The Midterm Review of Strategy 2020 (ADB, 2014a) sets out a vision for mainstreaming adaptation and climate resilience in development planning, project design, and implementation. Subsequently, the ADB Strategy, Policy, and Review Department introduced mandatory screening of ADB infrastructure projects to identify those at high or medium risk of being adversely affected by climate change. At-risk projects must be "climate-proofed" to make them resilient to identified climate change impacts.
The process of adapting to climate change can be highly data-intensive (IPCC-TGICA 2007;Weaver et al. 2013;Wilby et al. 2009). Contextual information is needed for preliminary risk screening and to identify socioeconomic factors that define adaptive capacities. Site-level data are required to establish baseline conditions for referencing climate change impacts and avoided damages, as well as for project monitoring and evaluation. Climate change scenarios are used with sector-specific impacts models to quantify expected risks and to evaluate the efficacy of adaptation measures. Alternative adaptation options must be prioritized based on their benefits and costs, as well as other criteria agreed with project partners.
Limited availability of contextual and climate information can present significant obstacles to climate risk assessments (CRAs). This bottleneck is particularly acute in data-sparse regions where even information about past climate trends and impacts may be difficult to obtain. However, improved access to remotely sensed (satellite) products, global climate (gridded) data, reanalysis products, and platforms for accessing country data help to fill some information gaps.
This technical note provides a compendium of open access resources that could assist experts carrying out CRAs. Where feasible, identified resources are mapped to the information needs of specific tasks within the CRA process.

Aims and Overall Approach
Four sector groupings account for the majority of ADB adaptation finance. These are: (i) water and wastewater systems; (ii) energy, transport, and other built environment infrastructure; (iii) crop and food production; and (iv) coastal and riverine infrastructure.
With these sectors in mind, this technical note aims to provide: (i) tables of reputable sources of public data on climate, sea level, and historical climate-related disasters for the Asia and Pacific region, which could be accessed by experts carrying out CRAs (especially those working in data-sparse regions); and (ii) commentary on the points at which these data might be brought into the ADB Climate Risk Management Framework (CRMF) (ADB 2014b) and economic analysis (ADB 2015).
The above aims were addressed by systematically appraising the information needs of CRA experts following all 20 steps of the ADB CRMF. Potential sources of data were assigned to one of four categories: (i) contextual information about historic greenhouse gas (GHG) emissions, climate vulnerability, risks, and impacts; (ii) historic information on climate variability and change; (iii) projections of multidecadal climate change; (iv) information about future climate change impacts and major completed or ongoing adaptation projects.
In each case, a brief description is provided along with an outline of the advantages and disadvantages of the information.
This technical note begins with a list of some ADB resources that support CRAs, then an overview of the key entry points for data input within the ADB CRMF. These elements are followed by sections on each of the four categories of information, including example resources and their application. The note closes with a few suggestions for strengthening ADB's capacities in data retrieval and analysis.
ADB has developed a range of guidelines to raise awareness of climate hazards and to improve the climate resilience of projects (Table 1). The complete ADB archive is searchable by publication type, country, subject, language, and publication date. Climaterelated sector guidelines are available for agriculture, energy, transport, and water. Crosssector guidance is provided for economic analysis of adaptation options, strengthening resilience through social protection programs, evaluation of natural hazards, and disaster risk assessment.
In addition to the resources listed in

Data Requirements of the ADB Climate Risk Management Framework
The ADB CRMF helps to identify climate-related risks to investments in the early stages of project development, and to thereby incorporate cost-effective countermeasures in the final project design. This is a top-down, scenario-led, 20-step process with five phases of activity: (i) climate risk screening, (ii) climate risk assessment, (iii) adaptation assessment, (iv) implementation arrangements, and (v) monitoring and evaluation ( Figure 1). All phases require data with the exception of implementation arrangements (although even in this case, there may be planned activities to strengthen the information base as an outcome of the project). The following sections outline how the information requirements of these phases may be met (in part) by public data sources ( Climate risk assessment (steps 6 to 11). This phase requires project-specific data to determine the vulnerability of project components to changing climate conditions (step 6). However, national indicators (with their constituent scores) can help to situate the project alongside long-term socioeconomic drivers of climate vulnerability and readiness to adapt (e.g., Notre Dame Global Adaptation Initiative [ND-GAIN] Country Index, Appendix 1). Development of appropriate climate change scenarios can be extremely time-consuming and data-intensive depending on the level of detail needed to evaluate potential biophysical impacts and economic costs and benefits. Bespoke scenarios may be created for the project via a range of statistical downscaling techniques fit to available climatic data (Appendix 2). Alternatively, existing high-resolution climate change scenarios can be accessed via online weather generators (e.g., MarkSim), regional (e.g.,    Adaptation assessment (steps 12 to 16). The value added by this phase of activity largely rests on the quality of consultation and level of engagement with project partners when identifying potential adaptation options. Lessons about "successful adaptation" processes can be exchanged through knowledge sharing platforms, regional networks, and case study catalogs (e.g., Asia-Pacific Network for Global Change Research, Appendix 4). However, adaptation outcomes may not be realized for decades and, even then, reliable counterfactuals may be hard to find (Moser and Boykoff, 2013). Nonetheless, the ADB CRMF requires that incremental benefits and costs of various adaptation options are established through economic analysis (step 15). Hence, climate scenarios are needed to assess the economic consequences of the project with or without climate change and with or without adaptations.
Monitoring and evaluation (steps 19 to 20). Reliable baseline data and long-term, postproject monitoring are needed to evaluate investment impacts, outcomes, and outputs. Scope for developing meaningful performance indicators is partly constrained by available baseline socioeconomic and climate information (Appendixes 1 and 2). The recurrent costs of long-term data collection must also be considered when setting up monitoring and evaluation frameworks. Ultimately, the intention is to use such evidence to share learning about what mix of adaptation measures work best, where, and when. Monitoring may also be needed to trigger measures at various points along planned adaptation pathways (e.g., rising sea levels as a stimulus for progressive raising and/or relocating of port infrastructure). Haasnoot et al. (2013) and Ranger et al. (2013) have other examples of adaptation pathways.
Historic Greenhouse Gas Emissions, Climate Vulnerability, Risks, and Impacts Appendix 1 provides various sources of contextual information, including for existing climate-related hazards, disaster impacts, development indicators, socioeconomic vulnerability, and readiness to adapt to climate change.
Many of the global databases listed provide national annual statistics; hence, any regional and/or sub-annual variations can be obscured. For example, the Emergency Events Database (EM-DAT) is a global archive of mass disasters since 1900, searchable by country, cause, and impact ( Figure 2). National summary sheets and dashboards of climate risks are also widely available (e.g., World Bank Climate Change Knowledge Portal), but generic outputs should always be sanity-checked using local knowledge. In Figure 3, Tajikistan is entirely landlocked, so cannot be at a "very low" risk of coastal flooding, whereas Metropolitan Manila has more than a "very low" risk of coastal flooding. Other sources give data on individual disaster losses and areas affected (e.g., DesInventar Sedai, or Dartmouth Flood Observatory), but the precise definition of events may not always be clear.    data are available in a range of formats that can be interrogated by GIS to extract local information relevant to the project. Historic data-whether on the properties of glaciers or climate-related hazards (such as fires and dust storms)-provide baselines for evaluating future threats to a project.  Appendix 2 lists sources of historic climate information for standard meteorological variables (such as temperature and precipitation over land areas), seasonal teleconnection patterns (such as for El Niño Southern Oscillation [ENSO]), remotely sensed variables (such as snow and ice cover), derived products (such as gridded weather data or reanalysis variables), and climate effects (on drought, food security, fire, and human health indicators). In each case, users should consider the temporal and spatial resolution of the various data sources in relation to the accuracy and precision required for the intended application.
Historic weather and climate data can be expensive to acquire or may be subject to strict licensing conditions. Sub-daily and daily hydrometeorological data are especially difficult to obtain from open sources and options may be particularly limited in data-sparse regions. The Global Summary of the Day (GSOD) is a searchable archive that provides access to daily variables for World Meteorological Organization (WMO) stations. However, records may be incomplete or contain unresolved quality issues so should always be checked carefully (Wilby et al. 2017). Sources of sub-daily weather data for selected sites in the Asia and Pacific region include RCCAP, Planet OS, and Weather Underground.
Other sub-daily and daily precipitation estimates can be obtained via the Giovanni portal for instruments, such as the Tropical Rainfall Monitoring Mission (TRMM). However, TRMM and other satellite-based precipitation estimates may be unreliable in mountainous, snow-covered regions, or near large water bodies, and where there are few land-based data for calibration (Karaseva et al. 2012;Yong et al. 2012;Dixon and Wilby 2016). A comprehensive evaluation of 30 global precipitation products showed that satellites tend to underestimate monsoon precipitation at high elevations in South Asia, and struggle to replicate observed frequencies of light and heavy rainfall in East Asia (Sun et al. 2018). Maps from the study show regional variations in systematic and random errors for satellite data.
Proxy daily weather variables can also be sourced from global reanalysis products, such as National Centers for Environmental Prediction (NCEP). These are essentially observed data from land, ocean, and satellite observations that have been assimilated by a climate model to create global coverage on a regular grid. Raw data are typically in NetCDF format, but some portals issue selected variables as TXT files from 1940s onward (e.g., SDSM).

Historic Weather and Climate Information
Monthly precipitation and temperature series are readily available from several sources (e.g., Earth System Research Laboratory [ESRL], and The Royal Netherlands Meteorological Institute [KNMI] Climate Explorer). The CRU Google Earth Interface is an interactive tool for locating nearest weather stations or interpolated data at 0.5° grid resolution for the period 1901-2014 (as in Figure 7). Although the underlying station data have been quality assured, reliability of the interpolated estimates decreases in data-sparse and complex terrain (Harris et al. 2014). Accuracy and coverage also diminishes for earlier decades, with much lower station densities before the 1950s (Figure 8). More recent political instability and conflict can lead to gaps in data too (   The KNMI Climate Explorer gives access to a variety of gridded, monthly weather variables from land-based observations and reanalysis products (Table 3). The portal can aggregate data by country, point location, or user-specified domain and season. Output is available as TXT files, time-series plots, or maps (showing change between chosen periods, or as anomalies and averages over time, Figure 9). When selecting data, one product might be preferred over another depending on the amount of interpolation applied, spatial resolution or period covered.

GISTEMP 1200
Relatively low-resolution (1,200 km interpolated) temperature data set, more accurate over large spatial (more than 1,000 km) and temporal (annual and longer) scales  Other sources of monthly climate and environmental series include river flow, drought, and water balance indexes (e.g., CAWater-Info, Global Climate Monitor, Palmer Drought Severity Index, Gravity Recovery and Climate Experiment); snow, ice, and glacier properties (e.g., National Snow and Ice Data Center); long-term sea levels (e.g., Permanent Service for Mean Sea Level); wind and sea surface temperature (SST) indexes (e.g., NOAA Climate Prediction Center). Global layers of monthly climate and bioclimatic 30-year means are available via WorldClim.
The ESRL correlation portal is a versatile tool for identifying significant teleconnections between various SST and atmospheric indexes, and remote responses in regional climate variables-the basis for seasonal forecasting. The tool is also useful for extracting and interpreting signatures of seasonal climate variability in long-term climate records. For example, Figure 10 shows the well-known teleconnection between Pacific SSTs in autumn and precipitation in winter across Southeast Asia. When the Niño3.4 index is strongly positive, there is increased likelihood of drought in the region. Some assert that developing a seasonal forecasting capability is one way of strengthening capacity to adapt to extreme weather both now and in the future (e.g., Washington et al. 2006).
The IRI Climate and Society Map Room is a library of graphics and data that monitors climate conditions at global and regional scales, along with metrics of climate effects on society (including food and water security, or human health). Many records are updated in near real-time and provide historic information, including trends and analysis of climate extremes. For example, Figure 11 shows the distribution of the 3-month Standardized Precipitation Index across Asia at the start of 2018 and the monthly time-series of this drought index for a site in India. Such information could be helpful as part of the preliminary evaluation of climate risks and in benchmarking climate change scenarios.

Multidecadal Climate Change Projections
Appendix 3 provides sources of information about future climate based on Global Climate Model (GCM) and downscaled regional climate scenarios. Even in an era of rapidly expanding climate services, access to high-resolution scenarios is more limited than might be expected. For instance, most portals offering daily climate model information (e.g., CORDEX, IPCC) are intended for use by the research community-with assumed expertise in handling NetCDF files or ability to assemble complex command sequences to access subsets of data from large archives. However, the most accessible data are only available as monthly, annual, or multidecadal means at the native grid-resolution of the GCM; downscaled and daily data are available, but in NetCDF format (CCKP) or only for Europe (KNMI). In Figure 12   The World Bank Water Anchor also produced scenarios (including mean air temperature, annual precipitation, a flood and a drought indicator) for major river basins. For example, Figure 14 shows climate changes (%) in the Chao Phraya river basin, Thailand by the 2050s based on CMIP3 output under SRES A2, A1b, and B1 emissions. However, the CCKP makes it clear that these data are not intended for use in any design study.
Source: World Bank Water Anchor. Climate model outputs are most readily available as monthly scenarios. This implies that their likely use within CRAs will be as change factors applied to higher resolution baseline data (IPCC-TGICA 2007). For example, the upper bound (97.5 percentile) ensemble projection for annual precipitation changes in the vicinity of Bangkok is 40%-50% ( Figure  12). In this case, a multiplier (change factor) of 1.4 to 1.5 would be applied to observed daily or sub-daily precipitation series for the city during the baseline period (1986)(1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005). Sensitivity testing of options within the economic analysis might scale across the ensemble range (0.9 to 1.5 for Bangkok) to represent the climate scenario uncertainty. Return period estimates can also be recomputed from the scaled series. However, in these types of application, a questionable assumption is typically made that all time-intervals (whether sub-daily, daily, monthly, or annual) scale by the same factor.
A limited amount of daily CMIP5 Regional Climate Model (RCM) output is available for Asia and the Pacific through the RCCAP portal. For example, the CSIRO RCM was used to downscale output from six CMIP5 GCMs (ACCESS1-0, CCSM4, CNRM-CM5, GFDL-CM3, MPI-ESM-LR, and NorESM1-M) to 10 km resolution for Thailand. Some of these scenarios were used to quantify crop water demand and rice yields in the Nam Oon Irrigation Project (Shrestha et al. 2017). Other RCMs, such as the PRECIS system, have been applied in the lower Mekong Basin (e.g., Mainuddin et al. 2013). A limited selection of outputs may be accessed via the Southeast Asia START Regional Center (SEA START RC) Data Distribution System and Analysis Tool (Figure 15).
Site-specific and bespoke climate change scenarios can be created by weather generators and statistical downscaling tools (IPCC-TGICA 2004). The latter may be calibrated using daily meteorological data from open sources, such as GSOD combined with reanalysis products from various portals. For example, the Statistical Downscaling Model (SDSM) has been applied in about 500 studies covering at least 25 countries in Asia and the Pacific region. The bibliography of SDSM research provides a starting point for developing high-resolution climate change scenarios for a range of sectors. Although the majority of studies focus on scenario creation for impact assessment, the latest version of the tool was intentionally designed to stress-test adaptation options (Wilby et al. 2014). The tool mm = millimeter, PRECIS = Providing Regional Climates for Impacts Studies, SRES = Special Report on Emissions Scenarios.
Source: SEA START RC.  Online weather generators (e.g., MarkClim) may deliver site-specific simulations of daily variables, intended for agricultural and hydrological impact assessments. Both types of downscaled product offer high-resolution scenarios conditional on coarser resolution climate model experiments and choice of emissions scenario. However, when applying such climate change information, care should always be taken to avoid mistaking increased precision for accuracy. Just because a particular tool or data set offers daily (or even subdaily) climate scenarios at a point does not necessarily mean that the projected changes are credible (Racherla et al. 2012). Ideally, indicators of climate model skill-that are relevant to the intended application-are evaluated as part of the scenario development process (e.g., Ekström et al. 2018;Wilby 2010).
Appendix 4 lists sources of information on climate change impacts (mainly for the water sector), programs of research and development with an adaptation dimension, as well as several knowledge platforms for exchanging intelligence and insights on adaptation (using case studies, project archives, and technical materials).
Some portals provide global risk assessments based on climate model scenarios and impact models. Archives of preformulated results are typically interrogated via mapping and visualization tools. For example, the Aqueduct Water Risk Atlas plots indicators of future water supply and demand under contrasting emissions scenarios ( Figure 17). Likewise, the DARA Climate Vulnerability Monitor (2012) gives national indicators of climate impacts and economic costs for 2030 compared with 2010 ( Figure 18). Care must be taken when interpreting such output because much is based on macro impact models applied at coarse scales with unspecified uncertainties. The information may be helpful for rapid appraisal of key climate threats and policy development, but unsuitable for detailed project design.

RCP = Representative Concentration Pathways
Source: Aqueduct Water Risk Atlas. , World Bank CCKP) offer national "fact sheets," "dashboards," or "country profiles" of sector-specific climate vulnerabilities and impacts. For example, the country profile in Figure 19 shows sector-level vulnerability levels, climate-related economic costs, and mortality, as well as the size of the affected population for Tajikistan in 2010 and 2030. This highlights acute vulnerability of the country to drought, flood, and landslides. Loss of hydro-energy is evidently the most significant economic vulnerability.  Fact sheets are intended to raise awareness within government and civil society of the underlying risk factors that increase vulnerability to emergent climate threats. They also enable risk screening and comparative analysis between different countries to inform investment decisions or resource allocation at programmatic levels. Indicative social and economic costs of climate impacts can demonstrate scope for avoided damages through adaptation investments. Some resources like the USAID Fact Sheets provide helpful links to documents on national strategies and plans, as well as tables of donor funded programs related to climate adaptation.
Knowledge platforms with searchable archives can widen access to information about adaptation projects across the Asia and Pacific region. Some specialize in key sectors (e.g., Asian Cities Climate Change Resilience Network) or regions (e.g., Climate Himalaya). However, web-based content can sometimes be dated or have broken links. Nonetheless, libraries enable shared learning from established impact case studies and adaptation projects. Such knowledge can inform the design, monitoring, and evaluation process of the ADB project implementation phase.
Source: DARA Climate Vulnerability Monitor.

Figure 19: Sample Country Profile of Tajikistan
This technical note is intended to support CRA experts, particularly those involved in the early stages of project development. Time and resources could be saved by attaching this document to terms of reference issued to CRA consultants. However, there is a limit to which globally accessible, open source data can meet the specific information needs of local adaptation projects. Hence, this note supplements rather than replaces efforts to gather relevant climate information from government agencies and counterparts, especially during the project concept phase.
This technical note compiles about 70 sources of public information for the Asia and Pacific region, including data on historical and future climate, climate-related disasters, indicators of national vulnerability, and preparedness to adapt. Additional sources will no doubt come to light so this is a living document-appendixes should be periodically refreshed as new information is located and quality assessed.
Most of the public data identified are contextual-providing high-level information about present and future climate risks at national and/or sector levels. Additional capacity development may be required in specialist "gatekeeping" skills, such as GIS, data homogeneity testing, post-processing research data formats (e.g., NetCDF), regional climate downscaling, and impact assessment. By growing these technical capabilities, more data could be accessed from the same public platforms or combined in ways that add value to the CRA.
As well as strengthening the technical capacities of local consultants and CRA experts, access to existing research-grade data stores should be improved through closer cooperation with scientific programs. For example, new interfaces could be developed to open research archives (e.g., CORDEX) to a broader user base. In particular, there is an urgent need to widen access to the daily and sub-daily climate information needed for economic analysis, engineering design standards, stress-testing adaptation options, and other aspects of project design.
Parts of the Asia and Pacific region are scarce of both the climate and socioeconomic information required for robust CRA. High elevation and remote environments are particularly problematic for data gathering, yet these locations have some of the most climate vulnerable communities and resources. Remotely-sensed and reanalysis products certainly improve coverage, but the accuracy of these assets ultimately depends on high-quality observing networks. Therefore, ADB, other multilateral development banks, and partner agencies should continue to invest in programs that strengthen national monitoring systems for climate and environmental change. Open access to such long-term records is invaluable for detecting emergent risks and devising, then implementing, effective adaptation measures.