Learning for an uncertain future Vulnerabilities of agricultural production to flood in the Sangkae river basin, Northwest Cambodia Jean-Christophe Diepart.

Learning for an uncertain futureVulnerabilities of agricultural production to flood in the Sangkae river basin, Northwest Cambodia

Jean-Christophe Diepart – Sotheavin Doch

Learning Institute, Cambodia

20.06.2013

Context

• Research program on Climate Change, Food Security and Natural Resources Management in Cambodia

• Eight research projects with 3 cross-cutting themes :

– Vulnerability

– Socio-economic differentiation

– Reconfiguration of State-Smallholders relations

Overview• Battambang, an agricultural

province

• Flood management in Tonle Sap plain is a long-time concern

• Flood as constraint and opportunity

• Future uncertainties

– Land use changes

– Rainfall pattern changes

→Objective: design/implement a methodology to understand the vulnerability of agricultural production to flood in main river basin

Theoretical framework

• Inadequacies of Risk-Hazards (RH) and Pressure-And-Release (PAR) Models

• “Global environmental” approach to vulnerability (Turner et al, 2003; Vogel et al, 2004)

• Vulnerability as the degree to which a coupled human-environmental system is influenced due to hazards and/or perturbation– Multi-dimensional– Cross-scale– Socially differentiated– Place-based– Dynamic

Human conditions

Environmental conditions

SensitivityExposure Resilience

Socio-ecological System

Coping

Impact

Adaptation (adjustment)

Adapted from Turner et al, 2004

Change in Human

condition

Interactions of hazards

Change in Environmental

condition

Human influences outside the place

Environmental influences outside the place

Vulnerability

Adaptation (adjustment)

Charact. of exp.

ImpactResponse

Place

World

Region

Methods

Multi-level vulnerability analysis

• Sangkae river watershed (3,708 km2)

• Commune

• Household

115 km

Methods

Watershed level→Correlation between

rainfall, forest cover change and main river water level

Methods

Watershed level→Analysis of recent flood

patterns based on participatory mapping– Minor (2010) and Major

(2011) flood– 3 types of flood (central

area, river overflow and run-off)

Methods

Commune level (n=31)→Indicator-based

vulnerability assessment

→Primary input data: commune survey (incl. mapping)

→Secondary input data: commune socio-economic statistics

Major Flood Area Size

Minor Flood Area Size

Major Flood Area SizeWeighted by flood duration

% agricultural area in commune

% population involved in agriculture

% cultivated area during rainy period

Major Flood Area SizeWeighted by impact on production

Institutional capacity% population living above poverty Density of road networkLiteracy rate

Expo

sure

Sens

itivi

tyRe

silie

nce

Exposure index

Sensitivity index

Resilienceindex

Indicators Index

Functional Typology of Commune

Vulnerability

Typology

Hierarchical Cluster Analysis

Methods

Household level (n=192)→Villages [5] selected

according to commune vulnerability types

→Survey on production activities and “exp.-sens.-resil.” to flood

→Analysis of differential vulnerability according to type of production system

Preliminary findings: watershed level

Flood types

Agriculture flooded area (Km2) % of

increaseminor flood

major flood

Central areas (Tonle Sap) 188.45 249.94 24.6

River overflow + Run-Off 64.38 120.22 54.08

River overflow + Run-off floods are significant, though off the records

Variation between minor-major event is more important upland

Preliminary findings: watershed level

No clear correlation between deforestation and water level fluctuation

Better correlation with rainfall patterns

YearsForest Cover

Change

Water level (m.asl) of Steung Sangkae river

Mean Max Min

Range (Max-Min)

Mean St. Dev.

1997-2002 10.3% 8.2 13.4 4.9 2.27 1.43

2002-2006 9.7% 7.4 13.3 4.8 2.07 1.80

2006-2010 24.6% 7.8 13.7 5.1 2.23 1.55

2010-2012 NA 7.6 13.9 4.8 2.09 1.34

Preliminary findings: commune level

Preliminary findings: household level

Important differences of vulnerabilities within “region”

Drivers of vulnerability are mostly associated with socio-economic differentiation processes between households (different resources endowments, labor management and inclusions/exclusions processes)

Work in progress…

Conclusions and implications

• The one-size-fit-all vulnerability is not suitable (Hinkel, 2011)

• Vulnerability is produced through different scalar configurations of human-environment interactions

• Indicator-based approaches are useful to categorize vulnerability, not to understand it

• Science & Policy dialogue– Natural Disaster Management – Spatial Planning – Watershed Management

Horizontal and vertical

integration !