Hitayezu p 20150708_1730_upmc_jussieu_-_room_101_(building_14-24)

Analysis of Factors Shaping Small-scale Famers’ Perceptions About

Climate Change in South Africa: A Behavioral Approach

1

Patrick Hitayezu, Edilegnaw Wale, Gerard Ortmann

2

Introduction

Climate change (CC) poses a real threat to South African farmers’ livelihoods.

Yet, small-scale farmers hardly recognize the climate trends.

Assessment of factors shaping farmers’ perceptions is scarce.

This study appraises the current perceptions and analyses their socio-psychological, cultural and institutional determinants.

3

Conceptual framework

Behavioural model of climate risk perception and response

4

Study area

uMshwathi local municipality, KwaZulu-Natal Midlands, a hotspot of CC in South Africa.

152 randomly selected farmers were interviewed.

5

Data

Dependent variables

Variable name

Variable description Mean SD

PERCEPTION 1 = local climate is changing; 0 = otherwise 0.684 0.466

CCP1 PCA index from PC1 of CC perception (contributed 45% to total variation in data)

-0.433 1.355

CCP2 PCA index from PC2 of CC perception (explained 24.7% of the variation in data)

-0.060 0.958

6

Data

Independent variables

Variable name Variable description Mean SD

AFFECT -2= CC is very bad/unpleasant/detrimental, … 2= very good/pleasant/advantageous

-0.101 0.051

KNOWLEDGE Number of correct responses about years with particularly abnormal rainfall

1.480 0.456

EGALITARIANISM Index of belief in human equality with respect to social, political and economic rights (α=0.81)

13.29 2.59

INDIVIDUALISM Index of belief in moral worth of an individual (α=0.73)

9.34 3.02

AGE Age of the household head in years (continuous) 58.940 12.83

GENDER 1 = Female-headed household; 0 = otherwise 0.532 0.400

EDUCATION Years spent by the household head in the formal education

6.552 3.951

: Aspects of vulnerability to climate change (sensitivity and adaptive capacity)

7

Data

Independent variables

Variable name Variable description Mean SD

EXTENSION Number of contact with extension workers in 2012 (count)

3.539 4.557

TRUST 1= don’t trust anyone, …, 4 = everyone is trustworthy 2.743 0.767

ADULTS Number of adult-equivalent members of the household

5.105 2.591

LAND Total operated area in hectares 1.596 1.515

RIVER Walking distance (in minutes) to the nearest river/dam

43.723 32.725

ROAD Minutes taken on arrive at the nearest tarmac road 12.559 17.774

AGRO-ECOLOGY 1 = Windy Hill Mistbelt (Mthuli); 0 = Wartburg/Fawnleas (Gcumisa)

0.322 0.468

: Aspects of vulnerability to climate change (sensitivity and adaptive capacity)

8

Actual climate trends

Annual rainfall (mm) and temperature (oC) records at Wartburg - Bruyns Hill Station (1972-2013)

A decreasing trend in rainfall is juxtaposed with a slightly increasing annual minimum temperature.

9

Actual climate trends Monthly rainfall (mm) records (1972-2013)

Decreasing summer rainfall and increasing winter rains

10

Actual climate trends Monthly min and max temperatures (oC) records (1972-2013)

Increasing minimum temperatures (particularly in winter)

11

Perceived climate trends

Unlike CCP1, CCP2 shows similarities with the actual trends.

Perception dimensions CCP1 CCP2 CCP3

Winter season is getting:

Abnormally colder 0.390 0.045 -0.297

Abnormally warmer 0.205 0.405 0.332

Abnormally dryer 0.315 0.169 0.107

Abnormally wetter 0.288 0.548 0.086

Summer (farming) season is getting:

Abnormally cooler 0.193 0.095 -0.104

Abnormally hotter 0.489 0.182 0.399

Abnormally dryer 0.852 0.390 -0.087

Abnormally wetter 0.019 -0.067 0.178

12

Results

Probit model results: Probability of perceiving abnormal climate

(showing variables with significant coefficients)

Variables Coefficient S.E. Marginal Effect P>|z|

AFFECT 1.025 (0.359) 0.201 0.005

EGALITARIANISM 0.445 (0.224) 0.095 0.046

AGE 0.238 (0.091) 0.051 0.013

GENDER 0.258 (0.111) 0.187 0.027

EDUCATION -0.056 (0.032) -0.022 0.082

AGRO-ECOLOGY 0.204 (0.075) 0.170 0.008

13

Results

Truncated model results: Strength of perceiving abnormal climate

(Showing variables with significant coefficients)

Variables CCP1 CCP2

Coefficient S.E. P>|z| Coefficient S.E. P>|z|

AFFECT 0.902 0.345 0.010 0.017 0.014 0.452

KNOWLEDGE 0.005 0.004 0.253 0.635 0.188 0.001

AGE 0.079 0.029 0.008 -0.183 0.152 0.494

EDUCATION 0.103 0.085 0.408 0.578 0.171 0.001

EXTENSION 0.132 0.083 0.174 0.133 0.066 0.048

TRUST 0.434 0.328 0.483 0.458 0.254 0.073

RIVER 0.053 0.018 0.005 0.012 0.008 0.198

14

Major findings

Congruent to the findings of other studies:

1. Affect is the most important predictor of the probability of perceiving climate risk among South African farmers.

2. The low probability of perceiving climate risk is not a result of knowledge deficit.

3. Perceptions about CC are deeply entrenched in farmers’ egalitarian values.

4. Women are more likely to perceive CC risk than men.

15

Major findings

5. Affect heuristics are the most important processes of ‘biased’ learning about CC among farmers (CCP1).

6. Knowledge (cognitive ability) is the most important predictor of the accuracy with which farmers perceiving climate risk.

7. Farmers’ investment in education and training increases their ability to receive and process climate information.

8. Farmers’ distrust in their community decreases their ability to perceive climate risk accurately.

16

Recommendations

CC information should relate to local farming realities.

Information on extreme weathers can be more persuasive.

CC communication should frame CC as a risk about which to worry.

CC communication should be well aligned to local beliefs, values and norms.

Awareness campaigns should involve affected farmers.

CC information should be communicated by locally trusted sources.

17

THANK YOU!

Funded by