Climate information: Does dissemination channels matter? Analysis of the coffee agroforestry system in the Sidama Region of Ethiopia

Abstract

This study analysed farmers’ access to and use of current information and communication technology (ICT) and non-ICT climatic information dissemination channels in the coffee agroforestry system of Sidama Regional State, Ethiopia. A cross-sectional survey design was employed. A multistage sampling procedure was used to select 360 coffee-growing farmers randomly. Data were collected using a semi-structured interview schedule. Descriptive statistics and the binary probit regression model were employed for data analysis. The finding shows that age, sex of household head, extension education and advisory services to climate change, membership in farmers’ groups, and mobile phone calls increased the likelihood of access to climate information dissemination channels. The marginal effect estimates show that the use of extension education and advisory services, access to television (TV), and access to frequency modulation (FM) handset radio increased the use of climate of information dissemination channels. Therefore, weather forecasts should be at a regional level through local media in local languages, and agricultural extension service delivery should include climate information delivery as one of the major activities.

Keywords:

• climate information
• dissemination channels
• access and use
• coffee
• Ethiopia

1. Introduction

The negative effects of climate change are already identified in all sectors, but with more impacts on agriculture (Ekemini et al., 2019; Olorunfemi et al., 2020). Agricultural production is threatened by extreme weather events such as drought and floods, with adverse effects on the fertility of the soil and crop productivity (Ozioko et al., 2022). Climate information services are becoming important and gaining recognition as critical to farmers and other decision-makers to manage climate risks and adapt to changing climatic conditions (Hansen et al., 2019). Climate information will aid farmers to tactically plan and adopt farm operations that enhance their adaptive capacity in the event of adverse climatic conditions and risks (Partey et al., 2020). Hence, the place of climate information services in helping farmers, particularly small-scale farmers, cope with climate change cannot be overstated (Eta et al., 2022).

Ethiopia is Africa’s largest coffee producer and the fifth-world producer, with 60–67% export earnings (USDA, 2022). Coffee is Ethiopia’s most important cash crop, with more than 15 million people directly or indirectly depending on it for their livelihoods (Feleke, 2018; Tefera, 2016; USDA, 2022). Sidama is one of the major coffee-producing regional states in Ethiopia, supplying over 40% of washed coffee to the central market of the country. Coffee has been the major source of income for rural households in coffee-producing areas of the Sidama regional state, and its yields vary over time (USDA, 2019).

Climate change is one of the important challenges for coffee production due to high temperature and rainfall variability, resulting in almost half the yields of standard years (USDA, 2019).

The lack of accurate, user-tailored, and timely climate information is a major obstacle to effective adaptation to climate change, especially for smallholder coffee farmers.

In order to improve farmers’ ability to adapt, the government must improve their access to accurate and timely agro-meteorological forecasts. By improving agricultural extension services, this can be achieved (Eshetu et al., 2021).

These climate information services are more accessible to smallholder farmers if ICT tools support them. This climate information is communicated through various dissemination channels, including radio, television, newspapers, mobile phone calls, text messages, and agricultural extension agents.

Climate information and advisories are considered useful tools in influencing farmer decisions to achieve sustainable agricultural production in the face of climate change. Adoption of climate change adaptation strategies by smallholder farmers could be efficient and sustainable if climate information is used in a timely way (Carr & Onzere, 2018; Carr et al., 2020).

Therefore, the objective of this study is to assess farmers’ access to and use of current ICT and non-ICT climatic information dissemination channels and determine factors that enhance the accessibility of climate information in coffee agroforestry systems.

2. Methodology

This research was conducted in Sidama regional state, Ethiopia, from 2018 to 2020 under the third round of the Hawassa University thematic research project. Sidama has geographic coordinates of latitude, North: 5’45“and 6’45” and longitude, East: 38’and 39’. It has a total area of 10,000 km², of which 97.71% is land and 2.29% is covered by water (Yibrah & Berihe, 2017). Sidama was previously a part of the Southern Nations, Nationalities, and Peoples’ Region (SNNPR) at the Zonal Administration level. It became Ethiopia´s 10th regional state in a referendum conducted in November 2019.

This research employed a cross-sectional survey design. A multistage sampling procedure was employed. In the first stage, coffee-growing areas were mapped in the technology villages of Hawassa University, and this helped to extract the geographical coordinates of coffee farms that represent coffee-growing areas of the study area. Accordingly, areas with higher land coverage for coffee production and high coffee suppliers to the export market were identified. Data for coffee supply to the export market was taken from the Ethiopian commodity exchange (ECX) and the regional coffee and spice development and protection authority. As a result, four districts (Shebedido, Dale, Aleta-Wondo, and Bensa) were selected purposefully based on coffee production and high export market supply. In the second stage, all villages in each district were listed and categorised into three agroecological zones (Highland, Midland, and Lowland). In the third stage, one village was selected randomly from the list of villages in each of the three agroecological zones and in each of the four sampled districts, making a total of twelve sample villages.

In the fourth stage, a total of 400 sample respondents were selected based on the total households in each district. Many of these households share similar characteristics as far as livelihood strategies, ethnicity, and socioeconomic status are concerned. Based on the proportion of total households living in each district, the number of sample respondents from each district was calculated. Yamane’s (1973) formula was used to calculate the sample size in this case. The formula is therefore presented as follows:

$n=\frac{N}{1+N{\left(e\right)}^2}$

Where:

n = sample sizeN = total number of households$e$ = the level of precision (0.05)

For each district, the above formula was applied, and 123, 117, 92, and 68 sample sizes were calculated for Shebedino, Dale, Aleta Wondo, and Bensa districts, respectively, resulting in a total sample size of 400. Finally, household heads were randomly selected from each village using the registration book at each village extension agent’s office.

The primary quantitative data, such as demographic, socio-economic, and institutional characteristics of the respondents, access to and use of ICT-enabled climate information, and non-ICT climate dissemination channels, were collected from coffee-growing farm households using a semi-structured interview schedule. The collected data were coded, cleaned, and entered using the IBM SPSS statistical package version 26 and STATA software version 14 for further analysis. This research employed descriptive statistical analysis such as minimum, maximum, mean, standard deviation, percentage, and econometric analyses using the binary probit regression model.

The probit regression model is used to analyse binary or dichotomous outcome variables (Soc & Anderson, 2014). The two-stage probit model is used to estimate the determinants of binary outcomes. This study used a two-stage Probit model (Borjas & Glenn, 1994) to identify the determinants of access to and use of climate information dissemination channels. In the first stage, the model identified the determinants of access to any climate information dissemination channels, where 0 and 1 values represented households who have no access and have access to any climate information dissemination channels, respectively. In the second stage, the model identified the determinants of use of any climate information dissemination channels, where the household has no use of and use of any climate information dissemination channels were represented by 0 and 1 values, respectively.

The equation for the binary probit regression model is:

$E\left(Y|X\right)=P\left(Y=1|X\right)=\mathrm{\Phi }\left({\beta }_0+{\beta }_{1}X\right)+\in$

The equation for the Probit regression model with multiple regressors is:

$P\left(Y=1|X_{1,}{X}_{2,\dots ..,}{X}_k\right)=\mathrm{\Phi }\left({\beta }_0+{\beta }_1+X_1+{\beta }_2{X}_2+\dots +{\beta }_k{X}_k\right)+\in$

where:

− P is the probability of a response− Y is a binary dependent variable− X is a vector of independent variables- ${\beta }_0$ is the intercept- ${\beta }_1$ is the coefficient for the independent variable X- $\mathrm{\Phi }$ is the cumulative standard normal distribution function-$\in$ is the error term which is assumed to be normally distributed.

This equation models the probability of the dependent variable Y taking the value 1, given the values of the independent variables X. The probit link function is used to model the regression function when the dependent variable is binary (Hanck et al., 2018). The maximum likelihood-based approach is used for the parameter estimation.

3. Results and discussion

3.1. Socio-economic characteristics of the respondents

Out of the total 400 semi-structured interview questions distributed, 360 were fully answered, but 40 were incomplete, resulting in a response rate of approximately 90%.

The result of the study revealed that the mean age and mean educational level of participants are 46 years old and 5.7 years of schooling, respectively (Table 1). Furthermore, Table 1 shows that the average family size was 6.2. While general farming experience, coffee farming experience, and total farm size were 28 years, 24 years, and 1.08 hectares, respectively, The mean total annual income of the respondents was about 10,594.32 Ethiopian Birr (ETB), which is equivalent to 197 USD.

Table 1. Socio-economics characteristics of the respondents

Variable	N. Obs	Mean	Std. Dev.	Min	Max
Age (yrs)	360	46.01	11.98	20	85
Education (yrs)	360	5.72	3.69	0	16
Household Size (no)	360	6.22	2.01	2	13
Farm Experience (yrs)	360	27.95	12.30	4	70
Coffee Farm Exp. (yrs)	360	23.65	12.18	1	70
Total Farm size(ha)	360	1.08	0.97	0.125	8.35
Total annual Income (ETB)	360	10594.32	9605.25	400	51080

3.2. Institutional characteristics of the respondents

This research revealed that more than two-thirds (62.5%) of the respondents have received extension advice on climate-related topics, while 56.4% of farmers were members of the farmers’ service cooperatives (Table 2). These are promising activities of the district bureau of agriculture, and their membership in cooperative societies might be an opportunity for farmers to discuss with each other about climate adaptation strategies and earn climate information.

Table 2. Extension education advice & Cooperative membership

Attributes	Response	Percent
Provision of extension education advice on climate related issue	No	37.5
	Yes	62.5
Cooperative membership	No	43.6
	Yes	56.4

4. Access to ICT-enabled climate information dissemination channels

Table 3 shows that 63.9% of the farmers have access to mobile phones and SMS text, while 48.6% and 41.1% of respondents have access to FM handset radio and traditional radio, respectively. This implies that ICT is an important channel, and as a result, we must focus on how ICT-enabled information dissemination channels can be used to enhance the greater flow of climate information to empower smallholder farmers in decision-making, especially in climate change. A study by Feleke (2015) complements this study in that farmers with access to weather information most commonly use radio as the communication medium, except for a few model farmers who use both television and radio.

Table 3. Access to ICT-Enabled climate information dissemination channels

Dissemination channels	Access to ICT
	Yes (%)
TV	81(22.5)
FM handset radio	175(48.6)
Traditional radio	148(41.1)
Mobile phone	230(63.9)
SMS text	230(63.9)
Broacher	0 (0)*
Poster	0 (0)*
Newspaper	0 (0)*

*Don’t have access to the communication asset. Source: Owen survey 2019.

5. Use of ICT-enabled climate information dissemination channels

Table 4 revealed that, out of those who have access to communication assets, 48.6% used FM handset radio and 44.4% used mobile phones as their major sources of information about climate change on coffee production. The implication is that the continued access to and use of ICT tools like FM handset radio and mobile phones offer diverse opportunities for rural communities to use timely and relevant climate information to enhance their livelihood strategy (Yohannis et al., 2019).

Table 4. Use of ICT-Enabled climate information dissemination channels

Dissemination channels	Users	Non- users**
	Freq. (%)	Freq. (%)
TV	62(17.2)	19 (23.5)
FM handset radio	175 (48.6)	0* (0)
Traditional radio	108 (30.0)	40 (27.0)
Mobile phone	160 (44.4)	70 (30.4)
SMS text	14(3.8)	216(93.9)
Broacher	0 (0)*	360 (100)
Poster	0 (0)*	360 (100)
Newspaper	0 (0)*	360 (100)

*Don’t have access to the communication asset, ** Calculated out of those who have access to communication asset. Source: Owen survey 2019.

On the other hand, none of the farmers utilised print media like brochures, posters, and newspapers for assessing climatic information on coffee production. This might be due to the absence of print media like brochures, posters, and newspapers in the extension advisory system to disseminate climate-related information.

6. Non-ICT information dissemination channels and their use

Table 5 revealed that farmers used village leaders 87.2%, development agents 81.1%, neighbor-friend-relatives 76.4%, Indigenous knowledge forecasters 66.1%, and farmer groups 50.5% as their source of climate information. This shows that i) village leaders play an active role in raising community members’ awareness of climate change and implementing climate change adaptation strategies. Village assembly meetings and other community gatherings were the main communication channels commonly used by village leaders. ii) Extension officers are the ones playing a key role in the dissemination of climate information to farmers; iii) The overwhelming majority of farmers acquire climate information from neighbor-friend-relative; iv) The importance of traditional weather and climate forecasting; and v) Farmers groups play a significant role in sharing climate information within their informal social network. Therefore, to increase farmers’ access to and use of climate information, the government must invest in non-ICT climate information dissemination channels.

Table 5. Non-ICT information dissemination channels and their usage (N = 360)

Dissemination channels	Users	Non- users
	Freq. (%)	Freq. (%)
Village leaders	314 (87.2)	46 (12.8)
Development Agents	292 (81.1)	68 (18.9)
Neighbors-friends-relatives	275 (76.4)	85 (23.6)
Indigenous knowledge forecasters	238 (66.1)	122 (33.9)
Farmers group	182 (50.5)	178 (49.4)

7. Determinants of access to climate information dissemination channels

The result of the binary probit regression model analysis in Table 6 showed that age (0.0263), sex (0.974) the dummy value for male is 1, extension advisory service on climate change (0.405), membership in the Farmers’ Group (0.361), and receiving phone calls from the household increased the likelihood of accessing climate information. While an increase in the farming experience of the household head reduced the likelihood of access to climate information by −0.1%, This implies that, as the farming experience of the household head increases, the probability of accessing climate information decreases by 0.1%. This result suggests that more experienced farmers are less likely to seek or use climate information than farmers with less farming experience.

Table 6. Probit regression result of factors influencing access to climate information dissemination channels

Variables	Parameter estimation			Marginal effect (dy/dx)
	Coef.	Std. Err.	P>z
Age of household head	0.0263**	0.011	0.019	0.010
Sex of household head (*)	0.974**	0.491	0.048	0.367
Schooling years	−0.017	0.019	0.354	−0.007
Extension advisory service on climate change (*)	0.405***	0.146	0.006	0.157
Receive text message (*)	0.153	0.457	0.737	0.058
Ownership of TV (*)	−0.198	0.170	0.244	−0.077
Inclusion of FM handset radio App in mobile phones*	−0.009	0.142	0.946	−0.004
Membership in the Farmers’ group (*)	0.361**	0.148	0.015	0.139
Receiving phone calling*	0.666*	0.370	0.072	0.225
Total income	−1.31e-06	7.28e-06	0.857	−5.04e-07
Farming experience	−0.025**	0.010	0.018	−0.009
Constant	−1.494868	0.577	0.010
Probit regression Number of obs = 360 LR chi2(11) = 39.38 Prob > chi2 = 0.0000\ Log likelihood = −223.38364 Pseudo R2 = 0.0810				Marginal effects after probit y = Pr(ACCESSTOCLIMATEINFO) (predict) = .60267343

(*) dy/dx is for discrete change of dummy variable from 0 to 1; ***, **, and * sig at 1%, 5% and 10% level.

It should be noted that this finding may not be universal and can vary depending on the specific context, region, and characteristics of the farming population. However, it is possible that experienced farmers may rely on their own knowledge and expertise gained over time, leading them to perceive a reduced need for external climate information. They may have developed strategies or practices based on their experience that they believe are sufficient for managing climate-related risks.

8. Determinants of the use of climate information dissemination channels

The result of parameter estimation in Table 7 showed that extension education and advisory services (0.659), TV (0.563), and FM handset radio (0.672) increased the use climate of information dissemination channels. The marginal effect estimates also revealed that farmers who have received extension education and advisory services (21.3%) This implies a one-unit change in extension education and advisor services that raises the use of climate information by 21.3 percent, holding all other variables constant. TV (20.5%) and FM handset radio (22.9%) increased the likelihood of using climate information. This shows that farmers who have access to TV and FM handset radio are more likely to use climate information compared to those who do not have access to these media. This finding implies that the media can play an important role in facilitating access to climate information for farmers. TV and FM radio are widely available and accessible media in many regions, and they can provide timely and relevant climate information to farmers. For example, TV programmes and weather forecasts can inform farmers about upcoming weather patterns and climate-related risks, while FM radio broadcasts can provide updates on local weather conditions and climate-related events. It is worth noting that the quality and relevance of the climate information provided through these media can also influence farmers’ use of the information. Generally, it suggests that efforts to promote access to and use of climate information should consider leveraging existing media channels to reach a wider audience of farmers.

Table 7. Probit regression results of factors influencing the use of climate information dissemination channels

Variables	Parameter estimation			Marginal effect (dy/dx)
	Coef.	Std. Err.	P>z
Age of household head	0.001	0.006	0.873	0.000
Sex of household head (*)	0.362	0.588	0.537	0.111
Schooling years	−0.022	0.020	0.270	−0.008
Extension advisory service on climate change (*)	0.659***	0.165	0.000	0.213
Receive text message (*)	0.534	0.358	0.136	0.201
Use of TV (*)	0.563***	0.177	0.001	0.206
Use of FM handset radio (*)	0.672***	0.225	0.003	0.229
Use of traditional radio*	0.227	0.152	0.135	0.079
Use of mobile phone (*)	0.126	0.240	0.599	0.043
Neighbors as source of info (*)	−0.105	0.184	0.567	−0.037
Membership of farmers coop (*)	0.136	0.156	0.383	0.047
Constant	−1.887	0.672	0.005
Probit regression Number of obs = 360 LR chi2(12) = 65.37 Prob > chi2 = 0.0000 Log likelihood = −192.83778 Pseudo R2 = 0.1449				Marginal effects after probit y = Pr(USEDCLIINFO) (predict) = .29124524

(*) dy/dx is for discrete change of dummy variable from 0 to 1; ***, **, and * sig at 1%, 5% and 10% level.

9. Conclusions

In comparison to ICT-enabled information dissemination channels like FM handset radio (48.6%) and mobile phones (44.4%), coffee-growing farmers in the study area had greater access to and use of non-ICT information dissemination channels, such as village leaders (87.2%), development agents (81.1%), and neighbours, friends, and relatives (76.4%), for climate-related information.

To increase farmers’ access to and use of climate information, the government must invest in the climate information dissemination strategy by supporting local village leaders to deliver the national climate information strategy to their respective local contexts. The acknowledgment and understanding of the farmers’ group can also contribute to a more resilient climate change approach.

Therefore, based on the main results of this research, the following recommendations were made: (i) the provision of climate information should be a key component of agricultural advisory services; (ii) the public extension system should be improved through the use of ICT-enabled information dissemination channels in addition to the non-ICT information dissemination channels like village leaders, development agents, and farmer-to-farmer (neighbours-friends-relatives). For this to be accomplished, development agents must upgrade their knowledge and skills on climate change and variability, climate risk management, and the dissemination of climate information.

Competing  interests

The authors declare that they have no competing interest.

Authors detail

All of the authors are members of the faculty of environment, gender, and development studies in the college of agriculture at Hawassa University.

Ethics approval and consent to participants

The research involved human participants, who were informed about the purpose of the research and that their participation was voluntary.

Acknowledgements

The authors would like to thank the experts in the agriculture office of the district for their patience and support in getting the required supplementary data. Besides, the authors would like to thank the respondents for their dedicated willingness to participate in this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The work supported by Hawassa University Office of the Vice President for Research and Technology Transfer and Research programs (Third Round Thematic Research Fund).

Notes on contributors

Kacharo Deribe Kaske

Kacharo Deribe Kaske is the corresponding author of the paper. PhD in agricultural education and extension, MSc in Rural Development and Agricultural Extension (RDAE), BSc in Agricultural Extension (AgEx), Diploma in Crop Production and Protection Technology (CPPT)