Examining climate change adaptation measures: an early warning system in the Philippines

Humanity has adapted to its changing environment since the beginning of time. Although adaptation is not a new concept, its process has become complicated as various elements are introduced to and absorbed into systems. As societies progress, the factors affecting the way humans live multiply and intertwine with one another magnanimously. Consequently, humans' coping mechanisms in response to these changes have also become complex.

Societies have practiced various measures to adapt to the impacts of weather and climate. Some of these include crop diversification, irrigation, water management, disaster risk management, and insurance. However, studies on climate change adaptation only emerged when scholars began to understand climate change itself. For example, Butzer (1980 as cited in Smit and Wandel, 2006) considered “cultural adaptation” (i.e. human ingenuity including technological innovation and long‐range planning) in examining the predicted climate change and its expected impacts on world food supply. Since then, climate change adaptation has been studied for a variety of purposes (Smit and Wandel, 2006).

The current climate changes pose risks and consequences beyond what have been generally experienced, such as impacts related to frequent droughts or heat waves, accelerated glacier retreat, and intensified hurricanes (Adger et al., 2007). Accordingly, because the issue on climate change has only been intensely discussed and addressed recently, the appropriate and effective adaptation practices are not yet well established. Still, adaptation measures have been and are being implemented, albeit on a limited capacity. Public and private entities adopt these measures through policies and regulations, investments in infrastructure and technologies, and behavioral changes (Adger et al., 2007).

Climate change adaptation is primarily about the way people adjust to and trade‐off the costs and benefits caused by climate change. These changes are expected to disrupt the current existence of people, influencing decisions on where people live, how they conduct businesses, the types of work they engage in, and the patterns of goods and services consumption. Thus, adaptation can be a form of structural adjustment; it is the accumulated series of changes in decisions and actions by individuals (Productivity Commission, Australian Government, 2011).

However, adaptation is expansive. It involves the actions of both individuals and whole communities, and both the private (businesses) and public sectors (government). Thus, individual actions are just one aspect of a complicated line of connections among the various groups and sectors affected by climate change. Society's ability to adapt to the changing climatic conditions is also said to be a function of its wealth, education, institutional strength, and access to technology. The nature and level of society's development greatly determine the extent of its exposure to climate risks and its ability to adapt (Burton et al., 2006).

Societies adapt for various purposes. According to Smit et al. (as cited in Adger et al., 2005), many elements induce adaptation such as protecting the economic well‐being or improving people's and property's safety. Similarly, it can be exhibited in numerous ways, from market exchanges, extension of social networks, to actions of individuals and organizations in order to attain their own (individual) or collective objectives (Adger et al., 2005). For example, a farmer can adopt a new agricultural technology or a whole community can build new irrigation infrastructure to cope with drought. Thus, adaptation may be for an individual's own advantage or for a communal benefit.

Nevertheless, adaptation is constrained by institutional processes like regulatory structures, property rights, and social norms. Likewise, communities, cities, business firms, and markets all adapt within the limits of available technologies, regulatory systems and knowledge of present and future climate risks. Aside from the limitations, the success of an adaptation measure depends on how that action meets its own objective and how it affects the ability of others to meet their own adaptation goals. An individual or community may deem a strategy successful whereas others may not (Adger et al., 2005).

This intricate structure of climate change adaptation warrants an organized and systematic analysis, evaluation, and planning. It needs a framework that provides numerous types of analyses for various purposes. It also needs a model general enough to be applicable to various circumstances, yet dynamic enough to be modified in order to address the specific needs and characteristics of the system being analyzed. The latter condition is based on the assumption that each case or system is unique because of the various elements that compose a system. Therefore, the effect of climate change on a system cannot be generalized.

The climate change‐vulnerability‐risk model (CCVRM) developed by Cuevas (2011) has the potential to satisfy all these requirements. The model can act as an analytical guide in understanding and examining how climate change adaptive measures affect or change system capabilities. This model may also help evaluate a system's current conditions and accordingly help identify and plan the best adaptive measure to be used given a particular circumstance. Lastly, the CCVRM is a framework that can be used to examine the way systems react to climate change in order to minimize the casualties and costs brought by the phenomenon and to maximize the accompanying opportunities and possibilities.

To examine the linkages among climate change, vulnerability, and risks, the research is divided into three stages. The first phase^[1] involves developing the CCVRM. The second phase explores the application of the model as a tool for conducting a quantitative analysis of the effects of climate change. Particularly, this stage involves developing and computing significant indicators and indices that have been identified in the model. Lastly, the third stage involves performing econometric analyses on a specific case using the indices developed from the second phase.

According to Cuevas (2011), climate change adaptation involves three primary factors namely: the nature of climate change, the vulnerabilities of systems, and the risks these systems face. The nature of climate change is the externality that shocks a system; it is the stimulus that causes systems to be at risk. Climate change and the risks faced are linked by the system's inability (or ability) to absorb, withstand, or react to the stress introduced. These characteristics define the system's vulnerabilities. Meanwhile, vulnerabilities determine the magnitude of the risk encountered as externalities enter the system. Thus, vulnerabilities influence the impact of these risks.

Cuevas (2011) has established the linkages among these elements through the CCVRM, an analytical framework that can be used to methodically examine how climate change may affect systems. Specifically, the model can help examine the way systems react or adapt to the changes and then evaluate the effectiveness of these adaptive measures in addressing the risks they encounter. However, this model primarily concentrates on system vulnerabilities and not on the actual system risks.

This paper aims to illustrate how the CCVRM can be used as an analytical framework in a qualitative examination of a specific climate change adaptive measure. It also intends to determine the capacity of the model to provide a multi‐sectoral analysis of adaptive measures. The climate change‐vulnerability‐risk research is an on‐going study founded on the development of this model. Any succeeding studies on the subject are attempts to determine the model's potential to effectively assess climate change adaptive measures.

This study investigates the case of the community‐based early warning system (EWS) in Quezon Province, Philippines. The next sections of this paper will discuss the CCVRM and will describe some significant observed climate change events in the Philippines which lead to the case of Infanta and General Nakar municipalities. The succeeding sections will also examine the community‐based EWS of the municipalities through the CCVRM. Lastly, the summary and conclusions section will complete the paper.

The research of Cuevas (2011) is a three‐stage study. The first stage is the development of an analytical framework that exemplifies the interrelationships among the nature of climate change, system vulnerabilities, and system risks. The study assumes that the nature of climate change affects the system through its vulnerabilities, and in turn, these vulnerabilities create different types of risks. This stage has been completed resulting in the three‐dimensional CCRVM (Figure 1).

The model looks at the system on a macro level and applies a methodology that may be applicable to varying circumstances. However, as each system is unique; the way climate change would threaten a system cannot be simplified. Although the methodology that may be used with this model is universal, its application can be customized due to the model's dynamic design. Accordingly, the subsequent analysis may be adjusted based on the changes in any of the model's dimensions.

The CCRVM assumes that all systems are dynamic; change is the only constant element. The stress and shock levels that affect the system's vulnerability highly depend on the nature of climate change. The CCRVM identifies four nature of climate change, namely: variability, intensity, frequency, and quantity.

Variability covers all types of weather changes but focuses on the time periods of the occurrences. This nature deals with the rate of changes and encompasses the concept of time. Variability manifests through:

• wind speed (i.e. increase or decrease);

• annual differences in precipitation (i.e. rainfall, snowfall, etc.); and

• temperature (i.e. increase or decrease).

Intensity represents the strength of climate‐related occurrences (i.e. typhoons, cyclones, snowstorms, droughts, and heat waves), whereas frequency relates to the number of such incidents. Meanwhile, quantity deals with the amount, such as volume of rainfall. This nature illustrates the amount, either by the excess or lack of it.

One of the key elements of this framework is its emphasis on the nature of climate change, and not on the phenomenon itself. Each one or combinations of any of these factors are externalities that affect the different aspects of systems (i.e. economic, social, biophysical, or technological). They create stress and shocks and their effects highly depend on the extent of the system's vulnerabilities.

In the CCVRM, vulnerabilities characterize a system. Vulnerabilities determine how the system absorbs, withstands or reacts to the stress and shocks entering the system. A system may be vulnerable to:

• only one;

• a combination of; or

• all climate change occurrences, depending on its social, economic, biophysical, and technological characteristics.

Overall vulnerability, meanwhile, involves the combination of the individual vulnerabilities of systems.

The impacts of climate change depend on the system's vulnerabilities. Vulnerabilities determine the magnitude of the risk a system faces as externalities enter it. Vulnerabilities, therefore, link the nature of climate change to the risks created, and ultimately, define the impacts to the system.

Figure 1 shows how the CCVRM classifies the four types of system vulnerabilities (i.e. socio‐economic, technological, bio‐physical, and institutional) that stem from the elements defining a system.

Economic and social vulnerabilities are closely linked, thus, are combined in this model to form socio‐economic vulnerabilities. The chief elements in this vulnerability type are people and resources. Examples of indicators that describe this type are:

• income sources;

• economic well‐being;

• poverty status;

• demographic characteristics; and

• health and educational statuses of the people in the system, among others.

On the other hand, bio‐physical vulnerability is defined as those natural characteristics of the areas which can be affected by the hazards. It is determined by physical traits such as location, soil quality, vegetation, etc. It also includes the system's geographical features like the site of the place and its proximity to the hazard sources and events.

Technological vulnerability is characterized by the existence (or lack) of technologies and infrastructures that may strengthen (or weaken) the ability of the system to effectively adapt. This type of vulnerability is illustrated by the status (or absence) of warning systems, protective structures, crop breeding and irrigation, settlement and relocation or redesign, and flood control measures.

Lastly, institutional vulnerability focuses on the system's knowledge of climate change, enhancements and systems development. This involves the society's level of awareness and knowledge about climate change which can be indicated by the presence (or lack) of organized community or government programs to address the issue and the policies and regulations established to adapt to the phenomenon. Institutional vulnerability emerges from the existence of the climate change externality. It is absorbed by the system as an internal process because of the need to adapt.

CCVRM categorizes risks as:

• income;

• bio‐diversity;

• health;

• mortality; and

• infrastructure.

Income risk encompasses the economic effects of climate change. It is linked to social effects through poverty such that if people's income sources are disturbed, then their socio‐economic statuses will also be affected. Meanwhile, bio‐diversity risk is associated with the impacts of climate change on natural resources (i.e. ecosystems, plant and animal life, and the survival of species). Risks on human welfare involve health risks resulting from hazardous environment, water supply and food shortage, and extreme temperature conditions. Mortality risk, on one aspect, may be an extreme consequence from the health risks; on the other hand, it can be a direct effect from landslides and floods due to the increased levels of rainfall or forest fires caused by the sharp increase in temperature. Lastly, impacts on man‐made structures (i.e. buildings, bridges, dwellings, roads, power plants, and water reservoirs) significant to economies and societies comprise the infrastructure risks.

Figure 2 may also shows the possible relationships among the variables where, N_i represents the nature of climate change, V_j, the type of system vulnerability; and R_k, the kind of risk that may exist.

The N_i and V_j sides of Figure 2 works with the following assumptions:

• 1.

  Natures of climate change are externalities that enter the system. They are independent of the system and are not affected by the system's activities or characteristics. Thus, they may stand alone. They, however, may create changes to a system.

• 2.

  All types of vulnerabilities exist in the system; and a nature of climate change may influence a system through its vulnerabilities.

• 3.

  Relationships between a type of nature and a kind of vulnerability are exclusive. N₁'s association with V₁ varies from that of N₁ and V₂'s, and so on. N₁ and V₁ may have a relationship, and N₁ and V₂ may not.

• 4.

  Suppose V_j, is an index^[2]. In relation to N_i, it may be assumed to be 1≥V_j≥0. Vulnerability, in connection with a nature of climate change, increases as the value of V_j approaches 1. When V_j=1, the system is completely vulnerable to N_i. Conversely, the system becomes less affected by N_i as the value of V_j reaches 0. When V_j=0, the system is immune to N_i.

• 5.

  When V_j and N_i has no relationship, V_j=∅ (null).

• 6.

  Given the exclusive relationship between N_i and V_j, the following may exist:

• N₁'s linkage to V₁ is independent of N₁'s linkage to V₂.

• If N₁ is linked to V₁, then V₁≥0, in relation to N₁. If N₁ is not linked to V₁, then V₁=∅, in relation to N₁.

• V₁'s linkage to N₁ is independent of its linkage to N₂.

• It is possible that V₁≥0, in relation to N₁, but V₁=0 in relation to N₂.

• N₁ may be linked to all vulnerabilities, V₁, V₂, V₃, V₄.

• V₁ may be linked to all natures, N₁, N₂, N₃, N₄.

Meanwhile, the V_i and R_k sides of Figure 2 works with the following assumptions:

• R_k may exist depending on the value V_j.

• Suppose R_k is also an index^[2]. If V_j>0, then R_k>0. This implies that if the system is vulnerable to a nature of climate change, then it is at risk.

• If V_j=0, then R_k=0. This relationship indicates that the system is immune to a nature of climate change and therefore the system is not at risk.

• If V_j=∅, then R_k=∅. This suggests that the system vulnerability has no linkage to the nature of climate change, and this translates to the same relationship with the system risk.

• R_k has a direct correlation to V_j. If V_j increases, R_k rises and vice‐versa.

• Relationships between a kind of vulnerability and a type of risk are exclusive. V₁'s association with R₁ is different from that of V₁ and R₂'s, and so on. Similarly, R₁'s connection to V₁ is independent from the linkage between R₁ and V₂.

The linking factor between N_i and R_k is the vulnerability of systems, V_j. Therefore, climate change may only put systems at risk if N_i is associated with V_j and in this relationship, 1≥V_j>0.

To summarize:

• Case 1. N_i and V_j are not associated: N_i just enter but do not affect the system and V_j =∅ (null). No R_k associated with N_i.

• Case 2. N_i and V_j have a relationship and 1≥V_j>0: 1≥R_k>0.

• Case 3. N_i and V_j have a relationship and V_j=0: R_k=0.

The linkages among the three are further shown by Figure 3.

If Case 1 exists, then the top dimension of the CCVRM cube is independent from the rest. This signifies that the natures of climate change are not associated with the system vulnerabilities. In Case 2, the model links the nature of climate change to risks by identifying and describing the way each nature influences the system (through its vulnerabilities) and by determining and describing the risks resulting from the existing vulnerabilities. To illustrate, in an agricultural setting, let N₁ represent the quantity of rainfall, V₁, the socio‐economic vulnerability, and R₁, the income risk. In this example, the agriculture system is vulnerable to rise (or decline) in the “quantity” of rainfall. This poses a risk to the source of the people's livelihood, thus, the income risk. Based on Figure 3, this may be interpreted as, N₁ linked to V₁, consequently, affecting R₁ (N₁V₁R₁). Furthermore, let N₂ represent the increase in the “frequency” of typhoons. In this scenario, N₂ has a relationship with V₁, which may alter R₁ (N₂V₁R₁). Other combinations of N_i, V_j, and R_k are possible.

The degrees of the system's vulnerabilities determine the scale of the risks and impacts of the climate change externality. One nature may affect all vulnerabilities, or all natures may influence only one kind of vulnerability. Similarly, this pattern may exist between vulnerabilities and risks. The matrix can help analyze the different combinations of the variables, may it be an intra or inter dimensional examination.

CCVRM works under the principle that climate change is a temporal issue. Due to the nature of climate change, the types of climate‐related occurrences in the past may not be comparable to those occurring today. Likewise, they may further vary in the future. Thus, the analysis centers on the nature of climate change while the framework takes the time factor into account. Similarly, vulnerabilities are also dynamic. Through the adaptive measures employed, system vulnerabilities may change. A system may become impervious to a previous vulnerability, hence, may no longer be at risk to a climate change nature like variability. Although in some instances, it may develop new vulnerabilities to a new kind of externality.

For example, agriculture in an area may be less vulnerable to intensified rains through building better agricultural infrastructure or adopting application of new technologies. Even a simple change in cropping pattern, as an adaptive measure, may lessen the area's degree of socio‐economic vulnerability since this action will affect the income source in the system. A lower socio‐economic vulnerability may translate to a lower income risk. On the other hand, a system may become more vulnerable to extreme winds of cyclones/typhoons never experienced before, and thus, change the level/s of either technological or institutional vulnerabilities, or both. Because of the intensity nature of climate change (i.e. increased wind speeds), the previous degrees of system vulnerabilities may worsen, and consequently, aggravate the risks to which they are linked to. Thus, adaptive measures must address system vulnerabilities and not the risks encountered because vulnerabilities determine the risk to systems. This substantiates the linking function of vulnerability in the CCVRM.

This study analyzes a single adaptation program, specifically the community‐based EWS in the municipalities^[3] of Infanta and General Nakar, Philippines. It does not deal with an integrated analysis of different measures applied in the system due to limited time and resources. Although investigating a multi‐adaptation program is possible, this would have required more time and resources to implement. Likewise, only a simple evaluation based on model definitions is conducted, instead of a more extensive risk and vulnerability assessment.

An integrated analysis of adaptive measures would have produced a more complete vulnerability and risk assessment. Nevertheless, this study's approach illustrates the way a single adaptation program addresses the problems in the localities. It also identifies the other issues that need to be considered in order to minimize the localities' overall risks from climate change. Analysis was conducted through review of related literatures. No field work was carried out.

The Philippines is an archipelago in Southeast Asia, composed of about 7,100 islands. It is located along the typhoon belt on the North Pacific Basin in the Pacific where 75 percent of the typhoons originate (Polack et al., 2010). According to the National Statistical Coordination Board (NSCB, n.d.), the country historically (from 1947 to 2007) experiences tropical cyclones (TC) and typhoons^[4] (TY) most frequently in July (197 TC, 94 TY), August (189 TC, 88 TY), September (177 TC, 92 TY) and October (153 TC, 92 TY). Most TC and typhoons enter the Philippine area of responsibility (PAR) during the third quarter of the year, whereas the least number of cyclones and typhoons are experienced during the first quarter. Although the country experiences the highest number of typhoons during the third quarter, the strongest typhoons are recorded during the fourth quarter (Virola, 2008).

Philippine typhoons have been getting stronger since the 1990s. Recent typhoons have recorded greater amounts of rainfall and faster wind speeds. From 1947 to 1960, the strongest typhoon (typhoon Amy) that hit the Philippines was in December 1951, with highest wind speed of 240 kph. From 1961 to 1980, typhoon Sening (international name: Joan) occurred in October 1970 and registered the highest wind speed at 275 kph. The record holders in the next 20 years were Anding (international name: Irma) and Rosing (international name: Angela), both with 260 kph wind speed, which occurred in November 1981 and October‐November 1995, respectively.

However, from 2000 to 2007, all records were broken by typhoon Reming (international name: Durian) with a registered wind speed of 320 kph. In fact, the original typhoon categories or public storm warning signals (PSWS) in the country were only three, with signal number one (no. 1) being the weakest and signal number three (no. 3), the strongest. With the new record(s), the classification of typhoon intensity was modified to include a new category – typhoon signal number four (no. 4)^[5] (Virola, 2008; www.typhoon2000.com (accessed 3 January 2012)).

Aside from this, the typhoons' rainfall volumes have been increasing as well. In November 2004, Typhoon Winnie recorded around 342 mm of rainfall in nine hours, equivalent to a 15‐day rainfall for the month of November^[6] (David and Felizardo, 2006; DPWH, 2007). In November 2006, super typhoon Reming carried a 466 mm of rainfall together with its record‐breaking wind speed. This level was the highest recorded in 40 years (Inquirer.net, 29 September 2006). Three years later, in September 2009, the country experienced another destructive tropical typhoon, Ondoy (international name: Ketsana), which submerged several parts of the country's capital. Ondoy released 455 mm of rain for 24 hours straight on 26 September 2009. This quantity of rainfall is more than (almost one‐and‐a‐half times) the historical average for 1993‐2008 for the entire month of September (i.e. 364 mm). Moreover, this level of rainfall surpassed all averages for any (entire) month, except for July's 459 mm (Virola, 2009).

These events show that climate change is affecting the Philippines. The typhoons that hit the country have carried increased quantities of rainfall and intensified wind speeds. Consequently, these have brought devastating losses in the lives and livelihoods of the Filipinos, as well as losses to the local and national economy.

Typhoon Winnie in 29 November 2004 flooded towns in the province of Quezon. Specifically, the municipality of Infanta experienced floods reaching 12‐18 feet and submerged 35 of its 36 barangays^[7] (Crisostomo, 2007). Consequently, the main river that traverses the municipality, Agos River, overflowed by six meters and damaged the primary bridge connecting the town to the other parts of the province. Waters also breached the 300‐meter right bank retaining wall downstream, which worsened the inundation of the town (DPWH, 2007; IFNet, 2006). General Nakar in Quezon was also heavily affected by typhoon Winnie. Like Infanta, it is situated along the Agos River, thus, also experienced serious flooding. The numerous landslides caused by the heavy rains further aggravated the conditions in the two municipalities. Most of the landslides were of small size in volume, ranging from several m³ to hundreds m³, and had a shallow failure depth varying from less than 1 to 3 m. The deadliest landslide, however, had a failure volume of some 2×104 m³ and traveled far, resulting in 111 deaths in Infanta (Dai et al., 2007). Mortality from typhoon Winnie totaled to hundreds.

The areas had no food or potable water, electricity, and access to communication after the disaster. Medicines were also scarce during the next four days (Crisostomo, 2007). The succeeding typhoons that came in December 2004 further devastated the towns, and consequently, slowed their recovery. These series of events claimed more than 1,000 lives and displaced over 135,000 individuals (Dai et al., 2007; IFRC and ADRC, 2009).

To prevent massive losses from such kinds of disasters, the Social Action Center (SAC)^[8], in cooperation with the Christian Aid and the Department for International Development (DFID), developed the community‐based EWS project in Infanta and General Nakar municipalities in Quezon Province. The project installed a systematic communication and EWS in the areas. The system was designed to raise the awareness and develop the abilities of barangay officials and community leaders to respond to flood related disasters. Sections 3.2 and 4 will examine this adaptive measure using the CCVRM.

The CCRVM assumes that all systems are dynamic; change is the only constant element. The system's stress and shock levels, which affect the system's vulnerability, highly depend on the nature of climate change. As mentioned, the model is a three‐dimensional matrix with nature of climate change, system vulnerabilities, and risks as its chief dimensions. The CCRVM identifies four nature of climate change, namely: variability, intensity, frequency, and quantity. Meanwhile, the four system vulnerabilities are: socio‐economic, technological, bio‐physical, and institutional. Lastly, risks are defined as: income, bio‐diversity, health, mortality, and infrastructure.

Figure 4 shows the way CCVRM can be applied on the community‐based EWS. For this qualitative analysis, three risk categories are devised, namely: high, medium, and low. These classifications are directly related to the vulnerabilities associated with them. Hence, the classifications depend on the perceived vulnerability statuses. Risk is high when the risk‐vulnerability linkage is high and is deemed to affect the system substantially or at the maximum level. Risk is at medium when vulnerability is lower (as compared to high), thus, expected effects will also be lower. This risk level is considered under the assumption that greatest loss can be avoided, yet cannot be regarded to be at its lowest. It also represents a moderate linkage between vulnerability‐risk. Lastly, low risk is linked with low level of vulnerability, thus, would have minimum effects, harm, and damage.

Vulnerability assessments using this framework are based on the definitions provided in Section 2. They are described as the current state of affairs using the latest available data. These forms of assessment and definitions of vulnerabilities are designed in preparation for the second phase of the research wherein statistical indices will be devised^[9]. The community‐based EWS being implemented influence the levels of system vulnerabilities. Hence, such changes are incorporated in this model dimension. Risks are evaluated based on the vulnerability assessments. According to the assumptions in Section 2, alterations in the levels of vulnerabilities can result to changes in the degrees of risks.

Intensity and quantity – typhoons in the Philippines, in terms of maximum wind speeds, have been growing stronger in recent years. As presented in Section 3.1, the typhoon's amount of rainfall has also been increasing, sometimes in levels equivalent to a number of days (of rainfall) during the month of the said event. These changes, especially concerning the wind speeds associated with the typhoons, have led to the Philippine PSWS categories to be modified from three to four.

Table I summarizes the typhoon categories in the Philippines.

These two types of climate change nature (i.e. intensity and quantity), are perceived to affect the characteristics of the typhoons that hit the Philippines. However, since both are observed in a single event, they affect the vulnerabilities of systems simultaneously.

Meanwhile, the frequency nature of climate change is not considered to be significant in the analysis. As shown in Figure 5, the number of TC entering the PAR has consistently averaged around 20 every two decades from 1951 to 2010 (1951‐1970, 1970‐1990, 1991‐2010). Variability, on the other hand, may be represented by the rate of change in wind speeds of the strongest typhoons in the Philippines. Figure 6 shows that the estimated rates (of change) from one point (year) to the next, have no notable patterns (Figure 6).

The variability nature of climate change, however, may manifest in various forms. The quarterly frequencies of TC entering the PAR are also investigated to determine whether trends have varied over time. Most of the cyclones hit the Philippines during the third quarter of the year. This pattern still holds true from 1948 to 2009. Although some years have shown that most cyclones occur during the fourth quarter of the year, they can still be generally expected from July to September.

On the other hand, cyclones occur the least during the first quarter of the year (Figure 7). Recall that variability focuses on time periods of events and deals with the rate of change. Based on the preliminary investigation of this study, variability may not be linked with the system's vulnerabilities.

This generalization may be premature and needs to be further studied. Due to the limitations of this study, exhaustive research cannot be performed at this point. Thus, both the frequency and variability natures of climate change are assumed to be inapplicable in this CCVRM analysis.

The next section will present the relationship between the nature of climate change and the types of vulnerabilities present in the systems of Infanta and General Nakar. If there is a strong linkage between the nature of climate change and vulnerabilities, then the vulnerabilities will be based on the characteristics of Infanta and General Nakar that may be affected by the intensity and quantity natures (of climate change). If no relationship exists, the system's vulnerability descriptions will be based on the framework definition. These will be the qualitative representations of the quantitative analyses projected in the second phase of the climate change‐vulnerability‐risk research. This assessment will concentrate on the available indicators; thus, the extent of the analysis may be constrained by the limited statistics.

a1 and a2 – in 2009, poverty rate in Quezon Province is high at 32.5 percent, much higher than the national rate of 26.5 percent. This translates to 582,752 poor people in Quezon. Similarly, poverty rate among families is higher in this province at 24.5 percent, compared to the national figure of 20.9 percent. An estimated 98,426 families are considered to be poor^[10]. Meanwhile, the latest available poverty estimates on the municipal level put General Nakar among the poorest with 63.94 percent poverty incidence. Infanta recorded a much lower rate, although it is still considered to be serious at 32.38 percent^[10].

The Quezon residents' main source of livelihood lies on agriculture. The province's agricultural area is equivalent to 513,610 hectares or 59 percent of its total land area. The province produces various crops, including paddy rice (irrigated and non‐irrigated), upland rice, corn, banana and coffee. Quezon is primarily known for its coconut produce, being the largest producer in CALABARZON and the second largest in the whole country. The province is also a major producer of palay and gets most of its supply from the municipalities of Candelaria, Sariaya, Tiaong, Infanta, Pagbilao, San Antonio, Lopez and the Cities of Lucena and Tayabas.

Both Infanta and General Nakar have similar resource bases, including forest, copra, arable land, fertile soils, abundant fresh water, and fishing grounds. In Infanta, 33 out of 36 villages are engaged in rice farming and 11 are classified as fishing villages. Rice fields and roads leading to the villages surround the town. Infanta's local network is connected to the Famy‐Infanta road, with Infanta being the service, trade, and population centre of northeast Quezon Province. Meanwhile, 21 villages in General Nakar engage in rice farming whereas nine engage in fishing. The town's road network in the north of Poblacion (the town's urban area) is almost non‐existent for a number of kilometers. Unlike Infanta, economic activities in General Nakar are very limited^[11].

As of 2007, the population of General Nakar was at 24,895, whereas Infanta's was 60,346^[12]. These figures compose 5.2 percent of Quezon Province's 1.6 million population^[13]. The number of households in General Nakar and Infanta are 5,019 and 13,959, respectively.

Infanta is administratively divided into 36 barangays, with seven urban and 29 rural areas. General Nakar has only one urban area among its 19 barangays^[14].

b1 and b2 – Infanta and General Nakar, connected by the Agos Bridge, are situated along the Agos River, which has a drainage area of approximately 1,000 sqm. The towns are also located at the foot of the Sierra Madre Mountain. The banks are highly prone to erosion due to their high‐energy gradient during floods. From the Sierra Madre mountain range, water flows out to the Pacific Ocean.

General Nakar has the biggest land area (134,375 hectares) in Quezon Province, composing around 15 percent of its 870,660 hectares. Most of its land is composed on mountainous slopes, with some narrow coastal plains and river valleys. On the other hand, Infanta only comprises 1.5 percent of the province's land at 34,276 hectares^[15]. Its eastern and northern parts are low‐lying plains whereas the west and southwest areas are mountainous and hilly.

A large portion of General Nakar is composed of wood and forest land, occupying around 126,000 hectares of its total land area. Around 20 percent are primary forests located at mountain tops with steep slopes. The remaining 80 percent are classified as second growth forest.

Although wood and forestlands also exist in Infanta, it has a bigger mangrove area of about 3,000 hectares. It also has a coastal area of 490 hectares and a tidal flat equivalent to 3,548 hectares.

Due to the municipalities' proximity to running water, land composition includes clay, silt, sand, gravel, and other similar materials, creating an area of wide alluvial plain. The highest elevation is about ten meters, whereas the lowest is five meters above sea level.

Agos River has an erratic river course due to the heavy sedimentation and sharp changes in slope, as observed from the old river course. Thus, the current course may be temporary and can be altered with significant sediment load transport caused by heavy floods. Further downstream, a sandbar exists on the left bank whereas erosion has been observed on the right bank.

Quezon Province is at the southern part of the island of Luzon. Due to its geographical location, it is one of the provinces frequented by typhoons, along with the provinces of Samar, Leyte, and the Bataan group of islands, among others.

c1 and c2 – three water level measuring stations were built in Barangay Pagsangahan and Anoling in General Nakar and in Sitio Cacawayan, Barangay Magsaysay and Barangay Pinaglapatan in Infanta when the community‐based EWS was developed. These base stations are equipped with instruments that systematically measure the levels of water in the Agos River during typhoons or occurrence of extreme rainfall. Rain gauges have also been installed in the upper and lower portions of the Agos River.

Technology transfer, through techniques in hydrological data collection, analysis and other related scientific knowledge on climate forecasting, was also implemented in the communities. Aside from the water level, the following river conditions are also observed: water color, force of flow, debris movement and movement of current. These are accomplished through links with the different agencies like Philippine Atmospheric, Geophysical, and Astronomical Services Administration (PAGASA, n.d.)^[17], Manila Observatory and the National Institute of Geological Science of the University of the Philippines.

Communication technologies (i.e. electronic mail, facsimile machines, landline and mobile phones) have been adopted for a more efficient relay of information on weather status from the PAGASA to the various units in the warning system (Municipal Disaster Coordinating Council or MDCC, Barangay Disaster Coordinating Council or BDCC, Control Base Stations, etc.). Radio telecommunication technology (i.e. two‐way radios) has also been adopted for a more efficient (local) transmission of the monitored water levels in the Agos River.

Indigenous techniques, such as “batingaw or bandillo” (bell or town crier) as a local early warning device, are also used in this EWS.

Meanwhile, a 300‐meter cemented retaining wall surrounds parts of the town proper located along the river banks.

d1 and d2 – the EWS established a new practice of disaster preparation, which includes scientific observation and monitoring of events, use of communication technologies, and organized coordination of community groups. It also developed linkages among the local governments and organizations, national agencies, and the scientific communities. It has created new institutional arrangements while strengthening existing ones.

Series of actions have been implemented for a more efficient flow of operation. During periods of abnormal weather conditions, the water level measuring stations diligently observe the water levels. Then, the information gathered will be relayed to the base stations, who, in turn, will report the same to the intelligence disaster analysis (IDA) of the MDCC for data processing. Data will be analyzed and the situation will be evaluated to determine whether emergency measures should be implemented. Afterwards, the IDA transmits back to the base station the information on the immediate action needed by the communities. The base station then responds by setting the EWS for the communities at risk. When threat is imminent, the BDCC will warn the communities for a probable evacuation through the “batingaw and bandillo” (bell, town crier). The assigned groups and agencies will be assembled to assist in the evacuation process^[18].

To ensure the preparedness of these municipalities, the MDCC of Infanta and General Nakar underwent disaster risk management trainings. A search and rescue and medical first aid training was also conducted among volunteers from the MDCC and other members of the communities. Moreover, the project also reactivated the BDCC and conducted exercises to further develop their skills and capabilities in handling emergencies. Meanwhile, trainings on the use of radios were done to ensure an efficient communication scheme. Most importantly, volunteers, local organizations, the local government units, barangay leaders, and all the communities participate in implementing the EWS and in executing the emergency actions^[18].

The succeeding assessments represent the relationship between the systems' (Infanta and General Nakar) vulnerabilities and the risks identified in the model. A1 refers to income risk brought by the socio‐economic vulnerability. In this case, the source of income in these communities (i.e. agriculture) is one of the primary socio‐economic vulnerability characteristics linked to income risk. The agriculture sector suffers the most during these intensified typhoons events. Floods destroy crops and negatively affect the economic well being of the whole population. Thus, income risk, in connection with socio‐economic vulnerability, is deemed “high”.

Recall that “high risk” classification is defined as the condition with high vulnerability‐risk linkage that can affect the system substantially or at the maximum level. Since this is a single adaptation analysis, this analysis assumes that no other project addresses the possible effects of the intensified typhoons on agriculture. The same kind of interpretation applies to the other risk descriptions/assessments, such that B1 – the link between bio‐physical vulnerability and income risk; C1 – technological vulnerability and income risk; and D1 – the associations between institutional vulnerability and income risk.

• A1 – socio‐economic vulnerability and income risk linkage.

Risk level: high

Agriculture plays a very significant role in both the towns of Infanta and General Nakar, especially in the latter. Consequently, the sector is the major source of employment of most residents.

Flood incidents due to the large amount of rainfall from extreme typhoons greatly affect the residents' livelihood. Farms engaged in rice production are the most affected during these events. All the villages in General Nakar are engaged in rice farming whereas 33 out of 36 villages in Infanta are dependent on the activity. Backyard livestock and poultry breeding for local consumption breeding is the norm in these areas.

The high socio‐economic vulnerability of the area, as evidenced by the high poverty incidence, makes farmers' recovery from field and livestock destruction extremely difficult, particularly when a large proportion of the poor belongs to farming/agricultural families:

• B1 – bio‐physical vulnerability and income risk linkage.

Risk level: high

The location of the two municipalities, being situated along the Agos River, provides a constant threat to the income sources of the residents. Land in the area is prone to erosion; this increases the likelihood of flooding during a typhoon, thereby aggravating the damage to the area's agricultural production:

• C1 and D1 – technological vulnerability and income risk linkage (C1); institutional vulnerability and income risk linkage (D1).

Risk level: medium to high

The two towns have limited infrastructure designed to minimize the probability of inundation or lessen the water levels during floods. The towns also lack agricultural technologies, such as adopting crops with higher tolerance to excess water and other agricultural techniques suitable for flooded areas. Because of these, damage to agriculture is highly probable during typhoons.

However, the water level measuring stations installed and the flood forecasting techniques adopted can predict the occurrence of flooding and its magnitude. This set of information can be used to salvage parts of the agricultural produce from damage, thus, lessen the likelihood of a total destruction scenario.

The improved warning system gives people more time to prepare for the impending devastation. Moreover, the linkages established with the PAGASA and other weather stations improved the accuracy of information received and increased the reliability of forecasts. Accordingly, crops may be harvested earlier to reduce losses. Since water levels can already be predicted and information flow has become more efficient, residents are more capable to prepare, store crops into storage facilities, and avoid spoilage. Livestock can also be sheltered to lessen losses.

• A2 and D2 – socio‐economic vulnerability and biodiversity risk linkage (A2); institutional vulnerability and biodiversity risk linkage (D2)^[20].

Not applicable

• B2^[21] – bio‐physical vulnerability and biodiversity risk linkage.

Risk level: high

Part of the land cover (especially Infanta's), is composed of mangroves, which may be at risk during incidence of floods. However, mangrove in Infanta is estimated at only 8.8 percent of the total land area. The different species of trees and animals living in mangroves are threatened due to the danger faced by their home.

A greater part of the area is classified as forest due to the Sierra Madre mountain range, the “largest remaining tract of old‐growth tropical rainforest in the Philippines”^[22]. It is also the home of hundreds of wildlife species, a number of which are unique to the Philippines. Soil erosion and landslides caused by extreme quantity of rainfall endanger these resources:

• C2 – technological vulnerability and biodiversity risk linkage.

Risk level: high

The areas lack the technology or infrastructure that protects the mangroves, mountain slopes, and the rainforest from severe floods and landslides. Moreover, farmers in the area use environmentally harmful farming techniques such as the slash and burn technique and the conversion of the old forest to agricultural land. These practices, on top of heavy human encroachment and illegal logging, are some of the perceived factors that aggravated slope failure during the typhoon in 2004 (David and Felizardo, 2006).

• A3 – socio‐economic vulnerability and health risk linkage.

Risk level: high

The high number of poor families lessens the communities' ability to gather sufficient food and medicine supplies in times of disaster and isolation. In 2003, 16.9 percent of children aged zero‐seven years in Infanta were estimated to be moderately to severely malnourished, whereas General Nakar registered a higher percentage at 29.1 percent^[23]. These magnify the risk to the physical well‐being of the people in these towns, especially, among the youth.

As of 2010, Infanta has only five doctors, 25 nurses, and five midwives^[24]. The doctor to population ratio is roughly 1 to every 12,000. During disasters, medical help from within the community will not be available:

• B3 – bio‐physical vulnerability and health risk linkage.

Risk level: high

The soil quality and the overall topography of the area increase the likelihood of erosion during strong typhoons and heavy rainfalls. These events are typically accompanied by falling boulders, uprooted trees, and other sediment debris that may block both land and water transport paths. In 2004, Infanta and General Nakar were isolated, with bridges destroyed and roads blocked. The floating logs and driftwoods covered the coastlines and the river's entrance to the sea, preventing boats and ships to get into the towns.

Rescuers and relief assistance will have difficulties getting into the towns. Food and medical supplies, therefore, can be delayed, increasing the health risk of the residents.

• C3 – technological vulnerability and health risk linkage.

Risk level: high

There is a 300‐meter retaining wall at the right bank of the Agos River. However, the infrastructure was not strong enough during the 2004 typhoon such that water breached the same wall. As the walls gave in, water flooded the town proper.

Floods and prolonged stagnant water in the area due to lack of a drainage system endanger the population to malaria, typhoid fever, and other water‐borne diseases.

Data from 1990 to 2001 recorded that residents in General Nakar (46 percent) and Infanta (44 percent) use springs and manual pumps as sources of water. Infanta residents also depend highly on manual pumps (44 percent), although the town has the municipal system (28 percent) and deep wells (15 percent) as water sources^[25]. If flooding occurs, clean and readily available water for the population will be difficult to acquire since the sources will be affected:

• D3 – institutional vulnerability and health risk linkage.

Risk level: high

Volunteer groups and community members underwent search and rescue and first aid trainings to prepare the communities for natural disasters. Although this improved risk management is a significant feat, it only addresses the dangers faced by the population during the event itself and does not deal with the health concerns after.

• A4 – socio‐economic vulnerability and mortality risk linkage

Risk level: high

Around 80,000‐90,000 families may be affected by floods in the two municipalities. Given the level of poverty, the types of housing units are not capable of providing safe shelter. The structures of the housing units are not strong enough to withstand the intensity of strong typhoons. Risk on people's lives is high due to inadequate shelters.

Moreover, low lying areas are typically highly populated, compounding the threat from rising waters:

• B4 – bio‐physical vulnerability and mortality risk linkage.

Risk level: high

The location of the two municipalities provides a constant threat to communities. With soil erosion, the trees in the surrounding areas can be uprooted due to the strong flow of water coming from the upland areas and the river. Hence, landslides are most likely to occur, consequently increasing the risks to the lives of residents living on the slopes and at the foot of Sierra Madre mountain.

These incidents worsen the floods, as evidenced in the 2004 event. Logs, uprooted trees, and other debris carried by the run‐off water blocked the bridge connecting Infanta and General Nakar. All these characteristics increase the chance of people dying from the floods that may occur:

• C4 and D4 – technological vulnerability and mortality risk linkage (C4); institutional vulnerability and mortality risk linkage (D4).

Risk level: low to medium

Retaining walls at the river banks are not sufficient infrastructures to prevent Agos River from overflowing and to protect the populated areas in Infanta and General Nakar.

However, the EWS and the improved capacity to predict flood occurrences enable the residents to receive timely information and to become more alert to the impending disaster. It also offers more time to prepare for evacuation. The system, being a community effort, strengthened the coordination and cooperation among the residents toward a common goal and course of action.

Assigned groups and volunteers are trained to execute first aid or immediate medical treatment to injured people and perform rescue missions. Meanwhile, the improved risk management and the organized emergency response actions make the implementation of evacuation activities more efficient, thus, lessening the danger of losing lives.

The reduced levels of technological and institutional vulnerabilities developed the communities' ability to act effectively. This is evidenced by the lives saved during typhoon “Reming” in 2006 wherein affected families were rescued and evacuated to safety. Recall that this typhoon, to date, is the strongest typhoon that hit the Philippines, recording the most intense wind speed and highest volume of rainfall.

• A5 – socio‐economic vulnerability and infrastructure risk linkage.

Risk level: high

As of 2000, only about 59.1 and 53.6 percent of households in Infanta and General Nakar are structurally sound^[25] resulting from the high poverty incidence in the areas. Strong winds and continuous rain during typhoons would easily damage the houses in the area:

• B5 – bio‐physical vulnerability and infrastructure risk linkage.

Risk level: high

Aside from Agos River, there are a number of small rivers and creeks within the boundaries of the two towns. During extreme weather conditions, the towns are in constant danger from the water overflowing from these water systems, consequently putting the infrastructures within the vicinity at risk:

• C5 – technological vulnerability and infrastructure risk linkage.

Risk level: medium to high

Except for the concrete Agos Bridge, most bridges in Infanta and General Nakar are made of timber. Thus, infrastructures do not have the capacity to hold heavy materials and loads. The condition is magnified during typhoons since even the strong concrete bridges become impassable when water rises rapidly and debris from the mountains block the flow of water in the river, as experienced during the 2004 flood.

However, the improved equipment for measuring the water levels and for monitoring the amount of rainfall can lessen the damages and can give the residents more time to prepare for the impending devastation. Likewise, with this additional preparation time, structures or emergency equipment can be added to the existing infrastructures to strengthen them. The technology installed has presented the residents with the opportunity to prevent great damages from happening:

• D5 – institutional vulnerability and infrastructure risk linkage.

Risk level: high

The plans and disaster preparations established by the different institutions and groups in the areas concentrate on the arrangements for evacuation and on saving lives. Programs geared towards protecting infrastructures still need to be developed.

Overall risk: HIGH

The effective use of the model lies with understanding the linkages between the natures of climate change (a1, b1, c1, d1, a2, b2, c2, d2) and vulnerabilities in order to determine the levels of system risks (A1, B1, C1, D1, … , C5, D5). The qualitative analyses of system vulnerabilities concentrate on the characteristics of the systems which may be affected by the intensity and quantity natures of climate change. However, in some cases, they are just descriptions of the system vulnerabilities based on the framework definition. Risk assessments rely on these same vulnerabilities to establish the vulnerability‐risk relationship. This analysis, therefore, can also be shows by Figure 8.

The lower case letters represent the relationships between the identified natures of climate change (i.e. intensity and quantity), which influence Infanta's and General Nakar's the systems through their vulnerabilities. On the other hand, the upper case letters denote the linkages between the systems' vulnerabilities and risks. Vulnerabilities associated in the lower case letters are of the same kind as those incorporated in the upper cases.

Typhoons in Infanta and General Nakar municipalities in the province of Quezon, are common. However, recent events showed that the accompanying winds and rainfalls of these events have been intensifying. The floods caused by typhoon Winnie in 2004 were one of the most unforgettable calamities in these areas. It destroyed lands, infrastructures and natural resources, and cost people their lives. In response, a community‐based EWS project was implemented to minimize these kinds of devastation.

Based on the CCVRM analysis, the EWS adaptive measure primarily addresses the intensity and quantity natures of climate change. The intensified typhoons have linkages to all four vulnerabilities of the systems (i.e. Infanta and General Nakar municipalities). The levels of system vulnerabilities have changed when the EWS was implemented. The project's strategy, specifically the improved techniques and instruments for forecasting the weather and for measuring the water levels, reduces the technological vulnerability of the two towns. Consequently, more accurate and timely information on the typhoons' time of arrival and intensity are provided to the communities, local governments, and all other relevant units in the areas. This feat is also achieved through the improved linkages between the Infanta and General Nakar municipalities and the scientific communities like PAGASA.

Institutional developments and establishing sound arrangements are vital to make the EWS project successful. The community‐based design of the EWS signifies the importance of both the local authorities' and community members' participation. This significantly reduces the institutional vulnerability of Infanta and General Nakar. Given the changes in the levels of system vulnerabilities, the risk levels also change.

However, this analysis only focuses on the community‐based EWS project and does not include an integrated analysis of various adaptive measures. Thus, the changes in vulnerabilities brought by other programs are not captured. Since the social and biophysical vulnerabilities are not affected by the EWS, they remain constant. Thus, they are in the same state before and after the EWS project was implemented.

Figure 9 outlines the two municipalities' risk levels given the CCRVM vulnerability assessment. All the risk classifications are influenced by all types of system vulnerabilities, except for the bio‐diversity risk. This particular risk is primarily influenced by the bio‐physical and technological vulnerabilities. This situation is an example of Case 1 mentioned in Section 2.

This approach in risk assessment is most useful in the quantitative aspect of the CCRVM^[26], although it is still effective in qualitatively examining the relationships among the different dimensions of the model.

Based on the CCVRM assessment, the community‐based EWS has the greatest impact in reducing the mortality risk in the system. The EWS project reduces the systems' technological and institutional vulnerabilities from greater intensities of typhoons. Due to this, reduced mortality rate can be expected. However, because socio‐economic and bio‐physical vulnerabilities are high, the total mortality risk level is still considered low to medium.

Although this study shows the technological and institutional vulnerabilities in relation to mortality risk are low, this same relationship cannot be generalized for other vulnerability‐risk relationships. This is because each linkage (although related to the other linkages to some extent) is independent from one another. The previous discussion on Figures 2 and 3 show this argument. Thus, technological vulnerability can be assessed as “high” if it is connected to biodiversity and health risks, while it can be considered “medium” to “high” if associated with income and infrastructure risks.

Figure 9 shows that the EWS project responds mainly to the intensity and quantity natures of climate change. It further shows how vulnerabilities affect the level of risks encountered by systems. When the degree of vulnerability decreases, so does the level of risk. However, all types of vulnerabilities that may have a linkage with a kind of risk should be addressed. As such, using this process can help achieve the least possible risk level.

This case study also shows that even with low technological and institutional vulnerabilities, overall mortality risk is still low to medium due to the system's high degrees of socio‐economic and bio‐physical vulnerabilities. Thus, if an additional vulnerability will be low (i.e. socio‐economic vulnerability), then the overall degree of mortality risk will change and will be within the minimum range levels. Similarly, a change in the degree of any other vulnerability in the matrix will alter the overall level of the associated risk.

The paper demonstrates how the CCVRM can be used to analyze climate change adaptive measures. The CCVRM is applied to the community‐based EWS being implemented in the municipalities of Infanta and General Nakar in Quezon Province, Philippines. The EWS project was devised in response to the constant threat of flooding in the area, especially during typhoons.

The levels of technological and institutional vulnerabilities of the two systems have lessened when the EWS project began to be implemented. These have direct correlation with the mortality risk, thereby, lowering its degree. Although the reduced levels of technological and institutional vulnerabilities influence the other types of risks, the effects are not as strong as their effect on mortality risk.

The project, meanwhile, does not show any relationship with the other system vulnerabilities, namely, socio‐economic and biophysical. Thus, these two remain constant and their levels are the same before and after the EWS project was implemented.

The paper investigates the adaptive measure being implemented in the two localities. It also shows the CCVRM's capacity to analyze planned measures, provide information on the types of vulnerabilities that can be affected by such projects, and identify the possible types of risks that may be influenced in the process. Although the study only concentrates on a single project, the CCVRM can also be used in a multi‐program assessment. The dynamic design of the model can transform the complex investigation of various programs into a systematic and organized integrated analysis.

Sining Cuevas is currently a PhD Research Student at the School of Geography, Planning, and Environmental Management (GPEM), University of Queensland, Australia. She also has a Master of Arts in Economics degree from the School of Economics, University of the Philippines. She worked as a Consultant at the Economics and Research Department of the Asian Development Bank (ADB) and was involved in various projects dealing with database management, informal employment, and low carbon growth. Sining Cuevas can be contacted at: s.cuevas@uq.edu.au