**3. Role of early warning systems and adaptation planning within the context of disaster risk reduction and CRM**

#### *"Neither society nor the environment are static. Consequently, neither is the risk" (Alan Lavell)*

It is clear that CRM may encompass a wide range of temporal and spatial scales, depending on the nature of the risks and their socio-economic context. Both climate and disaster risk are considered as some integration of hazards, vulnerabilities and exposures (see Section 1), and EWSs may be integral to the adaptation measures used to manage these risks. An important concept for Disaster Risk Reduction (DRR) taken from ISDR Glossary, 200936 concerns "the risk of disaster", which is usually expressed as 'the probability of life loss or property destruction or damage in a given period of time'. The action of DRR usually refers to the socio-economic objective of reducing that risk. In comparison, CRM is focused more towards the longer-term application of climate information and tools in a multidisciplinary scientific context to address both the positive and negative impacts of climate variability and change on society, infrastructure and life. The concepts of DRR and CRM are therefore complementary, both including a focus on risk management, and there are mutual advantages involved in designing integrated DRR and CRM projects.

Improving Climate Risk Management at Local Level –

Techniques, Case Studies, Good Practices and Guidelines for World Meteorological Organization Members 493

vulnerability and governance39,40,22. Countries or regions that have developed such systems may also use them to develop and inform strategic adaptation response options to climate changes, thus developing broader institutional flexibility and preparedness, and reducing

For most locations early warning is still a linear process based on a "sender-receiver" model of risk communication. In this section, the term "early warning information system" is used to describe the more integrated process of risk assessment, communication and decision support, of which an "early warning" is a central output. An early warning information system involves much more than development and dissemination of a forecast, it is the systematic collection and analysis of relevant information about and coming from areas of impending risk that: (a) Informs the development of strategic responses to anticipate crises and crisis evolution; (b) Provides capabilities for generating problem-specific risk assessments and scenarios, and (c) Effectively communicates options to critical actors for the purposes of decision-making, preparedness and mitigation. Central to the implementation of this more comprehensive vision of "early warning information systems" is a detailed

Numerous international and national EWS' exist 42,43,44. In addition, many early warnings directly and indirectly activate other warning systems in affected sectors and communities, a process that has been referred to as a cascade of early warnings45. For the most part, EWS' have been interpreted narrowly as technological instruments for detecting and forecasting impending hazard events and for issuing alerts. This interpretation, however, does not clarify whether information about impending events is actually communicated and used to

An example of good practice with an EWS is provided by the Climate Forecast Application in Bangladesh (CFAB) project. Heavy rainfall episodes in the Ganges-Bhramaputra basin (combined drainage area ~1,662,000 km2) cause human suffering almost every year. Webster & Hoyos (2004)46 showed the possibility of using physically based statistical schemes to predict rainfall with lead times of more than 10-days in the monsoon region. Based on this and subsequent research the CFAB project, supported by Office of Foreign Disaster Assistance of the United States Agency for International Development (USAID/OFDA), was launched during the monsoon of 2003 and 2004. Long-lead forecasts for rainfall in the river basin were given using the UK-based European Centre for Medium Range Weather Forecasts (ECMWF), Tropical Rainfall Measuring Mission (TRMM) and other datasets47. The Program was a collaboration (see Fig. 2) of the following agencies: Atmospheric and Oceanic Sciences at the University of Colorado, Boulder, Georgia Institute of Technology, ECMWF, Bangladesh Meteorological Department (BMD), Bangladesh Flood Forecast and Warning Centre (FFWC), and Asian Disaster Preparedness Center (ADPC) in Thailand. Coordinated efforts by ADPC, BMD and the Institute for Water Modeling (IWM) resulted in the development of 1-10 day discharge forecasts at major stations of two rivers (Hardinge Bridge, on the Ganges and Bahadurabad, on the Bhramaputra). The FFWC was responsible to produce local-level forecasts in other locations along these rivers. The Center for Environmental and Geographic Information Services (CEGIS)48 disseminated flood forecasts

examination of the root causes of the lack of early action41.

societal vulnerability.

reduce risks44,22,46.

If the integration requirements of DRR and CRM are considered according to impact scales (local, regional or global), it is clear that even if some climate phenomena relate to global causes they are materialized through locally-specific contexts, causing damage or losses depending on existing capacities in those local areas. Hence, the need to focus on improving the management capabilities of both DRR and CRM is especially important at local scales. Early Warning Systems (EWS) are often central to DRR and CRM, particularly over relatively short time scales (minutes to weeks), whereas other systems and tools that focus on longer-term adaptation planning tend to be more appropriate for CRM at longer time scales (months to decades). All such measures have the common aim of reducing vulnerability, increasing resilience and improving response capacities of people, economies and ecosystems at risk.

The UN International Strategy for Disaster Reduction (UNISDR) defines early warning as "the provision of timely and effective information, through identified institutions, that allows individuals exposed to a hazard to take action to avoid or reduce their risk and prepare for effective response". Governments often maintain EWS to warn their citizens and themselves about impending hazards, resulting for example, from health, geologic, or, climate and weather-related drivers. Traditional assumptions are that effective functioning of EWS requires only prior knowledge of risks faced by communities and other users of the early warning information. Under a CRM framework EWSs are expanded to other adaptation planning measures, including technical monitoring and warning services for highlighting the risks and their potential impacts, effective strategies for dissemination of understandable warnings to those at risk, and finally, knowledge and preparedness to act37. Two additional elements have been introduced, 1) awareness that risks are changing (and which new risks may arise) and, 2) the need for constructing and communicating new knowledge about future conditions that can be understood, trusted and used38,22. One goal is to be prepared to use windows of opportunity for engaging and providing leadership, and for legitimizing risk management and successful communities of practice that have arisen during but also between events.

Given the links between near- and long-term climate variability and change, the early warning construct also applies to more extended timescales. For example, WMO 'Climate Watch' systems utilise near real-time and historic climate observations with proactive mechanisms for interacting between users and NMHSs to provide alerts on major climate anomalies and extremes (see http://www.wmo.int/pages/themes/climate/climate\_ watch.php). Improving the institutional organization of the EWS as well as the associated strategic response to crises are closely linked to developments in understanding of climate vulnerability and governance39,40,22. Countries or regions that have developed such systems may also use them to develop and inform strategic adaptation response options to climate changes, thus developing broader institutional flexibility and preparedness, and reducing societal vulnerability.

492 Risk Management – Current Issues and Challenges

and ecosystems at risk.

during but also between events.

property destruction or damage in a given period of time'. The action of DRR usually refers to the socio-economic objective of reducing that risk. In comparison, CRM is focused more towards the longer-term application of climate information and tools in a multidisciplinary scientific context to address both the positive and negative impacts of climate variability and change on society, infrastructure and life. The concepts of DRR and CRM are therefore complementary, both including a focus on risk management, and there are mutual

If the integration requirements of DRR and CRM are considered according to impact scales (local, regional or global), it is clear that even if some climate phenomena relate to global causes they are materialized through locally-specific contexts, causing damage or losses depending on existing capacities in those local areas. Hence, the need to focus on improving the management capabilities of both DRR and CRM is especially important at local scales. Early Warning Systems (EWS) are often central to DRR and CRM, particularly over relatively short time scales (minutes to weeks), whereas other systems and tools that focus on longer-term adaptation planning tend to be more appropriate for CRM at longer time scales (months to decades). All such measures have the common aim of reducing vulnerability, increasing resilience and improving response capacities of people, economies

The UN International Strategy for Disaster Reduction (UNISDR) defines early warning as "the provision of timely and effective information, through identified institutions, that allows individuals exposed to a hazard to take action to avoid or reduce their risk and prepare for effective response". Governments often maintain EWS to warn their citizens and themselves about impending hazards, resulting for example, from health, geologic, or, climate and weather-related drivers. Traditional assumptions are that effective functioning of EWS requires only prior knowledge of risks faced by communities and other users of the early warning information. Under a CRM framework EWSs are expanded to other adaptation planning measures, including technical monitoring and warning services for highlighting the risks and their potential impacts, effective strategies for dissemination of understandable warnings to those at risk, and finally, knowledge and preparedness to act37. Two additional elements have been introduced, 1) awareness that risks are changing (and which new risks may arise) and, 2) the need for constructing and communicating new knowledge about future conditions that can be understood, trusted and used38,22. One goal is to be prepared to use windows of opportunity for engaging and providing leadership, and for legitimizing risk management and successful communities of practice that have arisen

Given the links between near- and long-term climate variability and change, the early warning construct also applies to more extended timescales. For example, WMO 'Climate Watch' systems utilise near real-time and historic climate observations with proactive mechanisms for interacting between users and NMHSs to provide alerts on major climate anomalies and extremes (see http://www.wmo.int/pages/themes/climate/climate\_ watch.php). Improving the institutional organization of the EWS as well as the associated strategic response to crises are closely linked to developments in understanding of climate

advantages involved in designing integrated DRR and CRM projects.

For most locations early warning is still a linear process based on a "sender-receiver" model of risk communication. In this section, the term "early warning information system" is used to describe the more integrated process of risk assessment, communication and decision support, of which an "early warning" is a central output. An early warning information system involves much more than development and dissemination of a forecast, it is the systematic collection and analysis of relevant information about and coming from areas of impending risk that: (a) Informs the development of strategic responses to anticipate crises and crisis evolution; (b) Provides capabilities for generating problem-specific risk assessments and scenarios, and (c) Effectively communicates options to critical actors for the purposes of decision-making, preparedness and mitigation. Central to the implementation of this more comprehensive vision of "early warning information systems" is a detailed examination of the root causes of the lack of early action41.

Numerous international and national EWS' exist 42,43,44. In addition, many early warnings directly and indirectly activate other warning systems in affected sectors and communities, a process that has been referred to as a cascade of early warnings45. For the most part, EWS' have been interpreted narrowly as technological instruments for detecting and forecasting impending hazard events and for issuing alerts. This interpretation, however, does not clarify whether information about impending events is actually communicated and used to reduce risks44,22,46.

An example of good practice with an EWS is provided by the Climate Forecast Application in Bangladesh (CFAB) project. Heavy rainfall episodes in the Ganges-Bhramaputra basin (combined drainage area ~1,662,000 km2) cause human suffering almost every year. Webster & Hoyos (2004)46 showed the possibility of using physically based statistical schemes to predict rainfall with lead times of more than 10-days in the monsoon region. Based on this and subsequent research the CFAB project, supported by Office of Foreign Disaster Assistance of the United States Agency for International Development (USAID/OFDA), was launched during the monsoon of 2003 and 2004. Long-lead forecasts for rainfall in the river basin were given using the UK-based European Centre for Medium Range Weather Forecasts (ECMWF), Tropical Rainfall Measuring Mission (TRMM) and other datasets47. The Program was a collaboration (see Fig. 2) of the following agencies: Atmospheric and Oceanic Sciences at the University of Colorado, Boulder, Georgia Institute of Technology, ECMWF, Bangladesh Meteorological Department (BMD), Bangladesh Flood Forecast and Warning Centre (FFWC), and Asian Disaster Preparedness Center (ADPC) in Thailand. Coordinated efforts by ADPC, BMD and the Institute for Water Modeling (IWM) resulted in the development of 1-10 day discharge forecasts at major stations of two rivers (Hardinge Bridge, on the Ganges and Bahadurabad, on the Bhramaputra). The FFWC was responsible to produce local-level forecasts in other locations along these rivers. The Center for Environmental and Geographic Information Services (CEGIS)48 disseminated flood forecasts

to communities during the monsoon season, working in close coordination with the Disaster Management Bureau (DMB) and the Department of Agricultural Extension (DAE). With the additional lead times and tailored warnings, community level flood risks were better managed. Communities were able to mobilize in advance (e.g. move livestock to higher ground, secure their fishing nets) in order to protect their livelihood assets. The project ran a second phase from 2006 to 2009 with support from USAID Bangladesh through CARE-Bangladesh. The objective was to transfer technology from the USA to Bangladesh and to build the capacity of national and local institutions for a sustainable end-to-end generation and application of flood forecast products in high-risk locations.

Improving Climate Risk Management at Local Level –

Techniques, Case Studies, Good Practices and Guidelines for World Meteorological Organization Members 495

communities, successful early warning information systems integrate "input" and "output" indicators. Input indicators include measures of production potential, including rainfall, soil conditions, heat and crop and livestock growth. Output indicators include nutritional indices, behavioral indicators, and signals of economic activity, that deal with the food, water and other supply situations or changes in demand that result from scarcity51. The timing and form of climatic information (including forecasts and projections), and access to trusted guidance to help interpret and implement the information and projections in decision-making processes may be more important to individual users than improved

Experience provided by the U.S. National Integrated Drought Information System (NIDIS) and the United States Agency for International Development (USAID) Famine Early Warning System (FEWSnet) drought early warning information systems developed in the

Frame the goals and objectives of international and country and intervention strategy

Specify of reliable information provided by forecasts, especially for key climate features

 Improve understanding of the modulation and combined impacts of interannual and decadal-scale variations on agricultural and meteorological drought duration and

 Place multiple indicators within a statistically consistent triggering framework-crosscorrelation among units for rapid transitions (e.g. climate and vegetation mapping)

 Develop risk and vulnerability profiles of drought-prone regions and locales including impact of climate change adaptation interventions on food and water availability,

 Develop indicators and methodologies to assess the risk to environmental services, value and costs of environmental degradation, and impacts of water and crop subsidies. Inventory and map local resource capabilities (infrastructure, personnel, and government/donor/ngo-supported services) available to complement food and water

Conduct gaming scenarios with planners and decision makers for selected past and

with scenarios of potential surprise and cumulative risks at each scale.

enable the flow of information among information system components.

The NIDIS and FEWSnet experience also provides a good example to demonstrate that successful drought early warning information systems have explicit foci on: (1) integrating

Improve understanding on whether and how best to use probabilistic information

Map decision-making processes and identify policies and practices that impeded or

USA has led to the following recommendations for developing EWS:

before critical thresholds are met from onset to severity.

from a securities perspective (water, food etc), e.g FEWSnet, NIDIS. Strengthen the scientific and monitoring foundations to support early warning.

reliability and forecast skill.

i.e. ENSO.

severity.

access, and use.

program operations.

projected events to:

Develop a Governance structure.

**Figure 2.** Institutional linkages for 1 to 10-day forecast of rainfall in Climate Forecast Application in Bangladesh (CFAB) project. Solid lines denote forecast/product flow and broken lines indicate coordination between the institutions: Asian Disaster Preparedness Centre (ADPC); University of Colorado; Georgia Tech – Earth Atmospheric Sciences (EAS); Bangladesh Meteorological Department (BMD); Flood Forecasting and Warning Centre (FFWC) of Bangladesh Water Development Board (BWDB), Disaster Management Bureau (DMB); Department of Agricultural Extension (DAE); Center for Environmental and Geographic Information Services (CEGIS); and CARE Bangladesh.

The disaster research and emergency management communities have shown that warnings of impending hazards need to be complemented by information on the risks actually posed by the hazards and likely strategies and pathways to mitigate the damage in the particular context in which they arise. Effective "early warning" thus implies information is introduced into an environment in which much about risk and vulnerability is assumed49. Vulnerability analysis provides a contextual basis for early warning by identifying structural, water, energy, and food insecurity attributable to disruption of primary means of access including informal community safety nets50. As is long-recognized by the disaster, food and water security communities, and more recently the climate adaptation research communities, successful early warning information systems integrate "input" and "output" indicators. Input indicators include measures of production potential, including rainfall, soil conditions, heat and crop and livestock growth. Output indicators include nutritional indices, behavioral indicators, and signals of economic activity, that deal with the food, water and other supply situations or changes in demand that result from scarcity51. The timing and form of climatic information (including forecasts and projections), and access to trusted guidance to help interpret and implement the information and projections in decision-making processes may be more important to individual users than improved reliability and forecast skill.

Experience provided by the U.S. National Integrated Drought Information System (NIDIS) and the United States Agency for International Development (USAID) Famine Early Warning System (FEWSnet) drought early warning information systems developed in the USA has led to the following recommendations for developing EWS:

Develop a Governance structure.

494 Risk Management – Current Issues and Challenges

Agro met translation

to communities during the monsoon season, working in close coordination with the Disaster Management Bureau (DMB) and the Department of Agricultural Extension (DAE). With the additional lead times and tailored warnings, community level flood risks were better managed. Communities were able to mobilize in advance (e.g. move livestock to higher ground, secure their fishing nets) in order to protect their livelihood assets. The project ran a second phase from 2006 to 2009 with support from USAID Bangladesh through CARE-Bangladesh. The objective was to transfer technology from the USA to Bangladesh and to build the capacity of national and local institutions for a sustainable end-to-end generation

**Figure 2.** Institutional linkages for 1 to 10-day forecast of rainfall in Climate Forecast Application in Bangladesh (CFAB) project. Solid lines denote forecast/product flow and broken lines indicate coordination between the institutions: Asian Disaster Preparedness Centre (ADPC); University of Colorado; Georgia Tech – Earth Atmospheric Sciences (EAS); Bangladesh Meteorological Department (BMD); Flood Forecasting and Warning Centre (FFWC) of Bangladesh Water Development Board (BWDB), Disaster Management Bureau (DMB); Department of Agricultural Extension (DAE); Center for

BMD

DMB, DAE

End users

Interpretation

ADPC Climate forecast

Communication

FFWC Discharge

IWM

translation

ADPC, CARE ,CEGIS

ADPC, CARE,CEGIS

Climate (rainfall and discharge) forecasting technology (EAS)/ADPC

The disaster research and emergency management communities have shown that warnings of impending hazards need to be complemented by information on the risks actually posed by the hazards and likely strategies and pathways to mitigate the damage in the particular context in which they arise. Effective "early warning" thus implies information is introduced into an environment in which much about risk and vulnerability is assumed49. Vulnerability analysis provides a contextual basis for early warning by identifying structural, water, energy, and food insecurity attributable to disruption of primary means of access including informal community safety nets50. As is long-recognized by the disaster, food and water security communities, and more recently the climate adaptation research

Environmental and Geographic Information Services (CEGIS); and CARE Bangladesh.

and application of flood forecast products in high-risk locations.

	- Improve understanding on whether and how best to use probabilistic information with scenarios of potential surprise and cumulative risks at each scale.
	- Map decision-making processes and identify policies and practices that impeded or enable the flow of information among information system components.

The NIDIS and FEWSnet experience also provides a good example to demonstrate that successful drought early warning information systems have explicit foci on: (1) integrating

social vulnerability indicators with physical variables across timescales, (2) embracing risk communication as an interactive social process and, (3) supporting governance of a collaborative framework for early warning across spatial scales46. Forecasts need not be perfect to make early warning useful. For longer-term EWS, it is also important to note that although a trend in the drought-based indicators may serve as a warning, the actual point of transition or threshold (e.g. dune mobilization) to increased severity remains difficult to predict.

Improving Climate Risk Management at Local Level –

Techniques, Case Studies, Good Practices and Guidelines for World Meteorological Organization Members 497

CRM is designed to be a practical process to be implemented on the ground. People, policies, environmental issues, governance, information, cultural aspects among other elements should be organised to communicate in an appropriate manner to deal with extreme weather- and climate-related risks. There is no single CRM solution for a particular situation in any part of the world. However, over time, there are increasing success stories demonstrating good practices for a wide range of CRM situations. The following case studies are examples of such good practice in CRM for different locations and development

**4.1. CASE STUDY 1: Using probabilistic seasonal forecasting to improve** 

**farmers' decision in Kaffrine, Senegal (Ousmane Ndiaye, Robert Zougmoré, Jim** 

Although agriculture and pastoralism occupy 80 per cent of the population in the Sahel, climate information is not yet widely integrated into farm management decision systems. However, many efforts have been made in the region to produce climate information such as the yearly climate outlook forum preceeding the incoming rainy season57. Yet, this hasn't benefitted the user community, particularly the most vulnerable to climate variability and change. This paper documents one ongoing demonstration project in Kaffrine, Senegal, within the peanut growing basin, where rural communities, policy-makers and relevant institutions are testing the use of probabilistic seasonal forecasts for managing climate risk. The process, from training the farming community to evaluating the use of the forecast

Rainfall in the Sahelian region of West Africa experiences strong variability over time-scales ranging from intra-seasonal (including long dry spells and false onset) to inter-annual and decadal. At the longest time scale, climate change is shifting the desert boundary and altering the landscape. This strong variability has an impact on many sectors, including health, agriculture and water management. The major impacts of climate variability in this region make CRM an imperative for the livelihoods of Sahelian communities. Each time

As is the case in most French colonized countries in Africa, ANACIM, the national weather service of Senegal, is in charge of providing meteorological services to the country. ANACIM, in partnership with the CGIAR research program on Climate Change, Agriculture and Food Security (CCAFS), initiated a pilot project in 2011 to test the usefulness of probabilistic seasonal forecast information to peanut farmers in Kaffrine. In addition to ANACIM, the key stakeholders participating in the project include local government technical services, local farmers and NGOs. A big challenge in the training was to go through many key and important steps in achieving good CRM, including producing

**4. Case studies demonstrating effective CRM practices** 

**Hansen, Aida Diongue, El Hadji Seck)** 

scale of variability requires a specific climate risk plan.

information, is outlined.

*4.1.1. Background* 

sectors.

Traditional warnings, with justification, remains an important source of climate information in many rural communities. At the community level, farmers in Zimbabwe and Malawi have identified local language radio programs as credible and accessible mechanisms to deliver forecasts if they occur with follow up meetings with extension agents or other intermediaries52. Internet based tools, such as Google maps, and graphical tools are already being used for participatory, large-scale information development. However, these tools are inherently limited in communicating the relevant local context and the consequences (positive and negative of information use). For most locations, the governance context in which EWSs are embedded is also key. The links between the community-based approach and the national and global EWSs are weak at present53. Improving the complementarity and legitimacy of both approaches is a new challenge to address especially in developing the institutional foundations for global climate early warning information systems envisioned by the Global Framework on Climate Services (see section 1. Introduction).

There is a critical need to approach and support early warning through DRR and Climate Change Adaptation (CCA)54,22, and the overarching processes involved in CRM. This requires a framework that uses climate change scenarios not above but within risk and vulnerability profiles, thereby capturing the nature of capabilities and decision-making networks. These form the basis for effective EWS design and implementation. The cases above, and other efforts, have demonstrated that social protection and early warning information interventions can provide DRR while helping to meet the goals of adaptation to changes in extreme events. Furthermore, sustainable development prospects are very dependent on the effectiveness of the many networks of EWS'57. In these networks, subtle rules of interaction emerge that shape the context in which resource-related decisions are taken, and the rules are negotiated and made55,56.

To ensure that DRR and CRM are integrated utilising appropriate systems, information and tools, some transversal capacities need to be established between the scientific community studying and analyzing the climate information (at timescales relevant to both DRR and CRM), and the decision-makers who are required to consider the full spectrum of the impacts of climate variability and change. Decision makers across all facets of society also need to be aware of the changes, risks and impacts threatening their societies and find appropriate ways to adapt to and protect these from the most damaging changes. They should also consider climate as a resource, with beneficial aspects that can be exploited, through application of timely and appropriate climate information, tools and products.
