World Resource Institute

Adaptation to Climate Change: Moving Beyond "Reactive" Approaches

By Shiv Someshwar, IRI/Earth Institute, Columbia University

Question One: Does climate change require new approaches to making decisions? Is the way we currently plan for the future and react to unexpected change sufficient to accommodate the uncertainty, scale, long lead time, and complexity associated with climate impacts?

Dramatic changes over the past several decades, particularly in developing countries, from land use changes, ecosystem modifications, and large-scale population movements, combined with weather/climate events beyond the "normal", wash away the risk calculation metrics that underpin development planning systems. Consequently, a better understanding of thresholds of risk in current systems, and an appreciation of the limits of buffering afforded to households and societies is badly needed. This will help to build anticipatory policies that allow decision-makers to manage dynamic systems.

To analyze limitations of current decision-making practices with respect to climate change, we first need to understand how institutions currently manage climate risks. While climate risks today clearly constrain economic growth and development, climate is already "hard-wired" into many processes, and economic development itself involves creating some capacity to buffer climate risks, while instituting robust mechanisms to respond to disasters. Adaptation to climate change requires moving beyond "reactive" approaches, toward finding ways to anticipate and manage risks. [1] This demands: a) better understanding of current and future risk thresholds; b) acceptance that some level of risk will remain to be managed "reactively," and c) an accounting for the dynamic nature of climate and non-climatic risks.

Examining the current workings of institutions dealing with climate-sensitive areas/decisions, three issues become clear: 1) climate is already "hardwired" or embedded in decision making; 2) current systems rely heavily upon reactive mechanisms; and 3) socio-economic systems are supported by formal weather/climate buffering arrangements.

1. Hardwiring of climate. There is a general appreciation of what constitutes weather/climate pattern, both temporally and spatially. Many countries follow the WMO suggestion of a 30 year period to define the climatology of a place/region. [2] Decisions on resource allocations and timing of decisions and activities are governed by this understanding of expected weather/climate. This is true at the micro-scale of farmer fields, in water supply management of mega urban regions, and at the macro-scale of national electricity distribution grids. Using history as a guide, decision making is optimized for weather/climate conditions within a "normalcy" bandwidth. Table 1 provides an example of decision making that is grounded in a shared sense of weather/climate conditions amongst paddy farmers of Indramayu district, in West Java, Indonesia. The activities shown involve the interactions among a range of institutions (that govern or help provide access to water, credit, labor, seeds, machinery, government programmed resources etc.) in order to prepare the fields and manage the several stages required for paddy cultivation. These interactions are based on an understanding of monsoon onset and structure (wet spell, dry spell, cessation), derived for the most part from shared experiential knowledge of weather/climate and within-season monitoring of actual rainfall. [3]

Table 1. Sample of agricultural activities and time line for paddy, Indramayu, Java, Indonesia

Meetings: Water user association + Irrigation Dept.; Farmer associations: labor / machinery contract, land preparation, variety selection


Seeds, financing mobilized (Bank, Coops); Seeds planted in nursery

Oct. 3rd wk

Drainage & Irrigation channels cleared

Nov 1st wk

On field, community-based and Water management

Oct - Jan

Seedling transplant

Nov 1st wk

Fertilizing , Weeding, Pest management, Labor contracts issued for harvesting, Harvesting undertaken, Post Harvest Storage / Surplus Disposal Decision, payment of dues"¦

Nov -- Mar


2. Reactive nature of current systems. When there is a departure in weather/climate conditions from expectations, institutions react to help mitigate the impacts. I use the 2005 Mumbai flooding experience in order to explore issues of "reactive management" in the absence of an overall climate risk management framing. [4] On July 26, 2005, Mumbai received an exceptionally large amount of rainfall (over 990 mm in a 24-hour period) with far-reaching human and socio-economic impacts. The city's drainage system was overwhelmed and resulted in the water-logging of several key transport arteries and the inundation of several suburbs under 15 feet of water. Over 630 lives were lost, with 50,000 residential buildings and 40,000 commercial establishments badly damaged. Since flooding is an annual ritual in Mumbai (though nothing close to this volume), the institutional mechanisms to deal with such disaster impacts were in place, and were largely responsible for averting an even larger human and economic catastrophe. [5] The same level of attention, unfortunately, was not paid in Mumbai (and in most cities of the developing world) to anticipating and planning for climate-related emergent risks. In the contentious days and weeks following the disaster, the direct causes of flooding were found to be have been long term in the making: an inadequate drainage system (designed in the mid 19th century) for a maximum rainfall of 25 mm/hour, large-scale land use changes that have resulted in severe reductions in surface water absorptive capacity and large increases in run-off, and encroachments into water drainage effectively reducing their drain capacity. The vulnerability of the system to flooding had been ratcheting up over the past several decades due to a lack of planning, poor infrastructure upgrading, mis-management and a general lack of awareness of the many drivers of vulnerability to flooding. The unusually heavy rainfall over a 24 hour period triggered a cataclysmic breakdown of the flood management system.

A key missing element in current climate response systems is the understanding of climate vulnerability as dynamic. In the context of socio-economic growth and climate variability and change, climate risks are not solely attributable to climate parameters. A singular focus on climate (as in many formal adaptation plans) misses the emergent nature of socio-economic risks. Dramatic changes over the past several decades, particularly in developing countries, from land use changes, ecosystem modifications, and large-scale population movements, combined with weather/climate events beyond the "normal", wash away the risk calculation metrics that underpin development planning systems.

3. Current buffering arrangements. Critical socio-economic systems are supported by formal weather/climate buffering arrangements. They include infrastructure, such as reservoirs, along with institutional arrangements for the timely distribution and supply of specific resources such as water, food, and electricity needed to soften the impacts of abnormal weather/climate events. This is true of farmers accessing common property resource (CPR) water supplies for supplementary irrigation. It is also the case for major reservoirs that can store enough water to suffice for 2 or 3 years, even with zero inflow (as opposed to run-of-the-river systems that are entirely dependent on current precipitation). This is akin to husbanding enough fodder (or food) needed to help tide over a potentially harsh winter season. In the short run, as long as weather/climate presents no major surprises (deviations beyond a certain range) most socio-economic systems (bar impoverished households and/or regions) can manage albeit at some cost. [6]

Anticipating and managing climate change risks:

Departures from historic climate averages, sustained by long-term anthropogenic impacts on the global climate system, pose a different set of management and response challenges to institutions. When return periods for drought are vastly reduced, the very structural characteristics the monsoonal systems are fundamentally altered, or the seasonal character of river flows from glacier melts is absent -- when the very climate characterizations that underpin institutional responses are genuinely altered -- this calls for a large-scale shift away from reactive climate management and towards anticipatory risk management. I turn next to three issues that would help societies better manage the potential negative impacts, and capture the opportunities from good climate conditions.

First, a better understanding of thresholds of risk in current systems, and an appreciation of the limits of buffering afforded to households and societies, is badly needed. This requires a more thorough understanding of the nature of variability (at weather and climate scales). Decisions and policies are typically based on a statistical average of past years. For example, the Colorado Water Compact of 1922 used average flows of the preceding 30 years to design water allocations (Chart 1). However, a longer view can help realize the limits of system resiliency afforded by this 30-year average. Projecting climate behavior into the future, using both potential low frequency variability behavior from the past and future anthropogenic impacts, can help decision-makers appreciate the potential range of climate variability and to understand the "resilient" bandwidth of the system.


Second, a politically-mediated process is needed to derive an "acceptable level" of climate resiliency of socio-economic systems. Using the best science to understand the impact of a changing climate is only one aspect, albeit a critical one. Equally important is to have a well-understood and politically inclusive process on determining the threshold of resilience to weather/climate hazards. Building "over-resiliency" is not an economic option, especially in LDCs and in the poorer regions of most developing countries. As a pragmatic alternative to climate "proofing," cities along the US Gulf Coast, for example, might plan resiliency to storms that are a Category 3 or lower. [7]Cities then would not try to build resiliency for higher category storms, instead opting for an emergency response system for such storms. Public expectation would be grounded in the possibility of being overwhelmed by a higher category storm. "Heading to the hills" could well be a rational social expectation, given the likely savings (especially to current tax payers). Elsewhere (Someshwar 2008), I have outlined some "climate-smart" development investment principles when analysis suggests future benefits from higher current investments. Those policies/decisions that provide some political cover ("if the expected disaster doesn't strike" or "is not as great as analysis suggested") are likely to be the ones prioritized by policy makers. They include for example, focusing on subsistence populations (whose "coping range" to climate shocks is non-existent), economically critical systems, extraordinary development opportunities (such as the development of a river basin), and piggybacking on infrastructural efforts underway (for example, the expansion of a metropolitan water supply sewerage system). The determination of resiliency thresholds depends very much on the capacity of societies to afford, implement and manage. However, this can be a thorny issue internationally, since relativity arguments could also be seen employing a "double standard" on human and social safety.

Lastly, in designing systems for managing impacts of climate change it is important to keep in mind that the non-climate shocks and/or trends are also as critical as the climate component. Similar to the case of flood management in Mumbai, for many existing climate sensitive systems, such as water supply systems, changing demands from population growth and higher levels of per capita demand often impose higher burdens (by orders of magnitude) than those due to changes in the climate. This is especially the case for the near term (out to about 30 years) in fast growing regions of the world (such as urban centers in Africa and Asia). Often, discussions of adaptation seem oblivious to the real world non-climate shocks that systems must respond and "adapt" to.

Anticipatory systems take time and effort to develop, and steps need to be taken now to advance this. By focusing on climate risks today, decision-makers have a more immediate interest in the outcome given the significant costs of floods, droughts and other disasters. This process will help decision-makers learn about managing inherently dynamic systems, and develop the system architecture for anticipatory action. Also, it would become evident that a singular focus on producing and making available better and more spatially nuanced information on place-based climate futures information would not automatically result in risk reductions. Also critically required are conceptualizations of risk in terms of costs and opportunities, consideration of climate risks in the context of a range of other risks, the importance of spatial, temporal and class dimensions of equity outcomes on system sustainability, as well as the design of institutional incentives for anticipatory actions. Only in this manner will societies develop the resilient institutions that are badly needed, as development risks from climate grow in the future.


Someshwar, S., 2008. "Adaptation as Climate Smart Development". Development. 51: 366-374.

Someshwar, S., E. Conrad, and M. Bhatt, June 2009. "From Reactive to Proactive Management of Urban Climate Risks in Asia: Institutional Challenges, Scientific Opportunities." World Bank Fifth Urban Research Symposium, Marseille.

Stahle, David W., Falko K. Fye, Edward R. Cook, and R. Daniel Griffin, 2007. "Tree-ring reconstructed megadroughts over North America since A.D. 1300." Climatic Change, 83:133""149.

[1] Adaptation in this note involves the development and implementation of well-resourced policies and decisions that use climate/environmental information and forecasts for equitable, efficient, effective, and participatory development in a manner that is sustainable now and in the future.

[2] Per WMO recommendations, a period of three decades is used as a standard period for characterizing the current climate.

[3] The climate information that underpins the entire ecosystem of paddy farming is an uneasy mix of indigenous understandings and more recent, computer-aided and government supported meteorological information. The potential power of the information (to access economic resources), more than its "goodness", seems to mediate acceptance and use of climate information locally.

[4] Parts of this discussion are drawn from Someshwar, Conrad and Bhatt 2009.

[5] The nodal Mumbai Disaster Management Committee, for example, had representation from a wide swath of government agencies including police, military, transportation, education, administration, food supplies, and meteorology, and was fairly successful in mitigating a higher lever of flood impact.

[6] The above discussion on current response management weather/climate should not be taken as a defense of current decision making arrangements. Far from it. Development is the best bulwark for climate resilience. Human mortality and morbidity to weather/climate is a simple and clear metric of "development". Within developing countries, the exposure of communities to climate hazards is further indication of their powerlessness.

[7] Discussion of "climate proofing", the rage of many multi-lateral institutions unfortunately lends the suggestion of being able to proof societies and communities against all weather/climate variations. This is fool hardy, even if economically possible. Raising the threshold of risk tolerance to a very high degree can, perversely, lead to very high impacts when such systems fail.