Effects of climate change-related hazards on food security and health within refugee settlements in Uganda

Abstract

Background

Climate change presents more dire impacts on food security and health in Africa than any other region worldwide. These impacts are expectedly more pronounced in predominantly agrarian refugee settings like Uganda, which is also the largest refugee host in Africa. However, there is limited contextual evidence regarding the effects of climate change-related hazards in Uganda to guide context-relevant interventions amidst the ongoing aid cuts. We assessed the impact of climate change-related hazards on food security and the health of refugees in three settlements in Uganda.

Methods

We conducted a cross-sectional study using concurrent mixed methods in Kiryandongo, Nakivale, and Rhino Camp refugee settlements. We collected data from 600 heads of refugee households using a structured questionnaire, conducted 45 in-depth interviews with refugees, and held 14 key informant interviews with program implementers. Quantitative data were analysed using STATA v16, and a modified Poisson regression model was employed to determine significant associations. Qualitative data were analysed inductively using thematic analysis to complement the quantitative findings.

Results

Of the 600 respondents, 60.6%, 95% CI = 56.5–64.6) were female, while most (76.2%) had lived in the settlement for 10–19 years. Climate change-related hazards associated with food insecurity included forest fires (aPR = 1.13; 95% CI = 1.08–1.23, p < 0.001), unreliable rainfall (aPR = 1.07; 95% CI = 1.01–1.15, p = 0.03), and landslides (aPR = 1.07; 95% CI = 1.01–1.14, p = 0.02). Qualitative findings highlighted that shifts in weather patterns, temperature fluctuations, and precipitation variations were perceived to have a profound impact on agricultural activities and to increase food insecurity and the incidence of vector-borne and waterborne diseases in the refugee settlements.

Conclusions

Climate change–related hazards can exacerbate food insecurity and contribute to health challenges by increasing the incidence of vector-borne and waterborne diseases. To enhance the resilience of refugee settlements against climate-related hazards, it is crucial to implement community-based disaster preparedness systems, invest in climate-resilient agriculture, upgrade water and sanitation infrastructure, and integrate climate risk assessments into public health planning.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13031-026-00817-x.

Introduction

Globally, food security is increasingly threatened by climate change and variability-related catastrophic events such as heavy rainfall, extreme temperatures, sea level rise and flash floods [1]. These extreme weather events disrupt food production, distribution and availability, potentially affecting food systems at every level of the supply chain [2, 3]. For vulnerable populations such as refugee communities in remote settings already grappling with poverty, political instability and limited resources, climate change and variability exacerbate existing vulnerabilities, heightening the risk of food insecurity and adverse health outcomes [4, 5]. Despite international commitments to attain Sustainable Development Goals (SDGs) 2 (to end hunger, achieve food security and improved nutrition and promote sustainable agriculture) and 3 (to ensure good health and well-being) by 2030, food insecurity remains a persistent challenge, particularly in refugee settlements (World Bank, 2024, UN, 2015).

The relationship between climate change and food insecurity is particularly pronounced in refugee communities. Often situated in ecologically vulnerable regions, these communities typically suffer from inadequate infrastructure, limited access to essential services, and a heavy dependence on humanitarian aid [6, 7]. Climate change-related events in these regions trigger a cascade of challenges, including water scarcity, diminished agricultural productivity, and heightened competition for limited resources [8, 9]. Such conditions not only exacerbate food insecurity but also increase susceptibility to environmental and public health risks, leading to increased morbidity and mortality [10, 11]. In addition, some studies have highlighted a rise in the incidence of animal diseases attributed to climate change and variability [12, 13]. In addition, the effects of climate change on crops and livestock are closely interconnected. For example, a decline in forage growth adversely impacts livestock weight gains [14, 15].

In African refugee settlements, food insecurity is escalating at an alarming rate, surpassing the challenges faced in other regions worldwide [16, 17]. This crisis can be attributed to the rapidly growing refugee population, which intensifies competition for food production and further strains existing resources [18]. Uganda currently hosts approximately 2.2 million refugees, and the number is constantly rising [19]. As a result, the country faces considerable challenges in combating food insecurity within the 13 refugee settlements. Nearly 90% of the refugee population lives below the poverty line, with the majority exposed to food insecurity regardless of how long they have lived in the settlements [20]. The ramifications of climate change extend beyond adverse health outcomes related to food insecurity to other ill-health conditions, such as an increased prevalence of vector-borne diseases, heat-related illnesses, respiratory conditions, injuries, and mental health issues, all likely to further burden the health system [21].

While existing evidence indicates that climate change intensifies food insecurity and its associated health outcomes [22–24], there is limited data on how specific climate change-related hazards impact food insecurity and health within refugee communities disproportionately affected by the hazards due to their limited coping capacity. This study examined the relationship between various climate change-related hazards and food insecurity, as well as associated health outcomes, in three refugee settlements located in different geographic and climatic regions of Uganda. The findings will inform interventions aimed at improving refugee food security and health, amidst the adverse effects of climate change and variability.

Methodology

Study site

This study was conducted in three refugee settlements in Uganda: Kiryandongo, Nakivale, and Rhino Camp. These settlements were selected from the thirteen refugee settlements across the country for their diverse climatic, geographical, ethnic and socio-economic contexts.

Nakivale, located in Isingiro district, Western Uganda, is one of the oldest refugee settlements in Africa. It hosts approximately 272,000 refugees and asylum seekers, as of 2025. It is geographically divided into three administrative zones: Base Camp, Juru and Rubondo, which further comprise 74 individual villages [19]. The population in the settlement is heterogeneous, comprising refugees from the Democratic Republic of the Congo, Rwanda, Burundi, and Eritrea. The refugees are granted access to land for farming and livelihoods, although the current agricultural methods are characterised by low yields, poor productivity and limited access to broader value chains.

Kiryandongo refugee settlement is located in Kiryandongo district, North-Western Uganda. It has a population of approximately 154,000 refugees and asylum seekers as of 2025. The majority of the refugees in this settlement are from South Sudan, with a small number from the Democratic Republic of Congo, Rwanda, Burundi, and Sudan [19]. The Kiryandongo settlement also serves as an interim stop for displaced people transiting to other camps. Up to 80% of the population are vulnerable women and children growing barely enough food to survive on one meal a day.

Rhino Camp refugee settlement is located in the West Nile region of Uganda. It is divided into 10 zones and 42 villages, covering approximately 225 km². It hosts over 125,000 refugees, primarily women and children. The majority of the refugees are from South Sudan. The primary economic activity for both refugees and host communities is agriculture. Due to significant underfunding of the World Food Programme, the refugee relief food rations in Rhino Refugee Settlement have been reduced by 40%, exacerbating chronic malnutrition among children in the settlement.

Study design

We conducted a concurrent mixed-methods study. The quantitative and qualitative data were collected and analysed simultaneously to complement each other [25].

Study population

For the quantitative component, we recruited refugee household heads who had been residents in the settlements for at least 10 years. A refugee household was identified as a family unit registered with the Office of the Prime Minister (OPM) and the United Nations High Commission for Refugees (UNHCR) as an independent home composition. Qualitatively, representatives from UNHCR, OPM, and implementing partner organisations with knowledge and experience in the study topic were identified for key informant interviews (KIIs). In-depth interviews (IDIs) were conducted with refugees who have lived in the settlement for at least 10 years, as well as with community leaders. These individuals were purposively selected for their knowledge of the community, long-term residence, and predominantly agrarian livelihoods.

Sample size determination

Yamane’s formula (1967) was used to estimate the sample size. Considering the total population across the three settlements - Nakivale (136,160), Kiryandongo (70,749) and Rhino Camp (121,580), a margin of error of 5% and a design effect of 1.5, we attained a sample size of 600 participants. We then applied stratified disproportionate distribution to allocate an equal sample of 200 participants across each stratum (refugee settlement), ensuring equal representation across settlements. However, while this approach ensured balanced representation across settlements, the study was powered for the aggregate sample, rather than settlement-specific estimates. The qualitative component of the study included 45 participants for IDIs and 14 KIIs. A minimum of 12 participants were targeted per settlement – a number sufficient to reach data saturation [26].

Sampling procedure

Participants for the quantitative component of the study were selected through a two-stage cluster sampling process. In the first stage, we obtained a list of villages within each settlement from OPM and UNHCR. We randomly selected five villages from each settlement on this list, resulting in a total of 15 villages. This selection was carried out using simple random sampling techniques: the names of the villages were written on pieces of paper, mixed, and then drawn without replacement. The choice to select five villages was based on logistical considerations, including resource constraints, the accessibility of the settlements, and the need to ensure geographic representation across each settlement. In the second stage, we identified households where the heads had resided in their respective villages for at least 10 years.

Based on the description in the sample size determination section, the 200 households per settlement were evenly distributed across the five selected villages, yielding 40 households per village. Within each village, we employed systematic sampling to select the households. If the household head was unavailable at the time of the interview, we proceeded to the next eligible household, maintaining the sampling interval until the target sample size was achieved. We excluded child-headed households and those who were ill.

For the qualitative components, IDI participants were purposively sampled based on their predominantly agrarian livelihoods, having lived significantly longer, a leadership role in the community in the settlement and willingness to participate. Conversely, the KIIs were selected for their knowledge and expertise in the study’s subject. Key informants included representatives from UNHCR implementing partner organisations and the Office of the Prime Minister.

Data collection

Research assistants with a bachelor’s degree in social and health sciences, as well as proficiency in English and the languages spoken in the refugee settlements (Kiswahili for Nakivale, Nuer, Dinka, Luo for Kiryandongo, Kakwa and Lugbara for Rhino settlement) were recruited and trained in interviewing techniques and research ethics. The study tools were pilot-tested in the Kyaka II refugee settlement, and we utilised the feedback to validate the variables and enhance the clarity of the questions before data collection commenced.

Quantitative data were collected in March 2023 using a structured questionnaire uploaded onto Kobo Collect-enabled Android mobile devices. The research assistants visited the selected households to screen potential participants for eligibility.

Qualitative data were collected through KIIs involving representatives from OPM, UNHCR and implementing partner organisations. All interviews were scheduled by phone and conducted face-to-face, except for the three KIIs, which were conducted virtually because the officers were off-site. These officers had a wealth of information and experience, and therefore could not be replaced by those on site. With the participants’ written informed consent, the interviews were audio recorded.

The interview topics were derived from findings of previous studies [24, 27], but also informed by the principal investigator’s knowledge and experience gained from past interactions with the refugee communities. However, we employed appropriate probing techniques to explore a wide range of information, leading to the emergence of themes and concepts relevant to the study objectives. These included types of climate change hazards, the impacts of climate change and variability on food security, and the burden of pests and diseases. We conducted interviews until data saturation was reached. On average, the IDIs lasted 42 min, and the KIIs lasted about 35 min.

Measurement of variables

The dependent variable for this study was food insecurity. Food insecurity was assessed using the Household Food Insecurity Access Scale (HFIAS) [28]. Participants were asked nine questions regarding their access to adequate food over the past 30 days. Each question warranted a yes/no response. A scoring system was developed for each household, categorising them into two main groups: “food secure” and “food insecure.” We combined those classified as food secure and those experiencing mild food insecurity into a single “food secure” category, while those experiencing moderate or severe food insecurity cases were grouped into the “food insecure” category [29]. This approach acknowledges that refugee settlements are typically food insecure; thus, mild food insecurity in the general population can be considered equivalent to food security within a refugee household.

Independent variables were informed by literature [30, 31] and anecdotal information from focal persons in refugee settlements. The categories of variables included social demographics, extreme weather events, and pests and diseases linked to climate change.

Data analysis

Quantitative data

We applied a sampling weight to account for the unequal probabilities of selection among participants across the three refugee settlements. This weight was calculated as the probability of selectingan individual aged 18 years or older in each settlement [32], relative to the total population across the three settlements. We used self-reported data to compute proportions that describe the sociodemographic characteristics of the participants and the distribution of various climate change events across the settlements. We did not aim to make settlement-specific comparisons, given that our sample size was not well-powered to detect these differences. Data were therefore analysed as an aggregate. Inferential analysis was performed using a modified Poisson regression model to determine which climate change-related events were associated with food insecurity within the refugee settlements. Data analyses were conducted using Stata version 16 software. To build the model, we manually employed both backward elimination and forward inclusion techniques, considering variables with p-values of 0.25 or lower. The final multivariable model yielded adjusted prevalence ratios with their corresponding 95% confidence intervals (CIs) at the 5% level of significance.

Qualitative data

Audio recordings were transcribed verbatim and analysed thematically. Two lead analysts (MTW and JN) individually read the transcripts and then collaboratively developed the coding scheme by reading the transcripts twice and taking notes on emerging themes and contentious findings. Through thematic analysis, codes were identified latently and organised into categories and ultimately into themes [33, 34]. The key issues related to climate change-related events and their impact on food insecurity and health were captured accordingly. We compiled our findings in a narrative analysis, accompanied by illustrative quotes. Two KII and three IDI participants provided feedback on the transcripts and preliminary findings. Qualitative data were managed and analysed using Atlas.ti version 9 software.

Results

We present the study findings by integrating the survey results with the qualitative evidence from the IDI participants and key informants, and also highlight areas of convergence in the reported climate change-related hazards across the three refugee settlements, We begin with a description of participant characteristics, followed by the primary study findings, which are organised into three broad qualitatively generated themes: climate change-related hazards, pest and disease burden, and effects on food availability, accessibility, utilisation, and stability.

Of the 600 survey participants, nearly two-thirds (60.6%, 95% CI = 56.5, 64.6) were female, and almost half (45.9%, 95% CI = 42.3, 51.4) had a household size of 7–12 members. Two-thirds (66.1%, n = 391) of the participants were married, while the majority, 80.5% (n = 468), had attained only a primary education level or lower. (Table 1).

Table 1.

Sociodemographic characteristics of the survey respondents

Variable		Frequency (n = 600)	Percentage (%)
Age-group (years)	18–30	171	25.4
	31–54	345	59.4
	55+	84	15.2
Gender	Male	228	39.4
	Female	372	60.6
Household size	1–6	242	43.5
	7–12	284	45.9
	13+	74	10.6
Years lived in settlement	10–19	457	76.2
	20–29	106	17.7
	30+	37	6.2
Marital status	Not married	209	33.9
	Married	391	66.1
Education level	Primary and lower	468	80.5
	Secondary and above	132	19.5
Major source of household income	Begging	9	2.0
	Remittances	245	34.5
	Salary/wages	30	5.1
	Non-agri-business	38	7.8
	Agriculture	272	49.5
	Others	6	1.1

For the qualitative component, more IDI participants were female (56.2%), while the KII participants were predominantly male (78.6%). The mean age of IDI participants was 34.5 years, and on average, they had lived in the settlement for 16 years (Table 2).

Table 2.

Characteristics of the qualitative participants

Characteristics of in-depth interview respondents
Age (years0	Average (34.5 ); Range (18–57 )
Sex	Female (56.2%); Male (43.8%)
Settlements	Rhino Camp (37.5%); Nakivale (31.3%); Kiryandongo (31.3%)
Number of years lived in the settlement	Average( 16 yrs) Range (11–20 years)
Household size	Average (8 people); Range (1–20 people)
Nationality	S.Sudanese (52.1%); Congolese(37.5%); Rwandese(4.2%); Burundian(4.2%); Kenyan (2.1%)
Highest level of education (formal)	None (6%); P1-P7 (31.3%); S1-S4 (35.4%); S5-S6 (6.3%); Bachelors (4.2%); Tertiary (10.1%)
Marital status	Single (33.3%), Married (60.4%), Divorced/ Separated (2.1%), Widowed (4.2%)

Characteristics for key informants
Sex	Females (21.4%); Males (78.6%)
Roles	OPM Representatives (3) UNHCR Representative (3) Health Implementing Partner (3) Livelihood Implementing Partner (3) Food and Nutrition Implementing Partner (3)

Climate change-related hazards across the three refugee settlements

Climate change-related hazards were reported across all three refugee settlements, although patterns varied. For instance, climate-change-related hazards were most commonly reported in Rhino Camp (46%), followed by Nakivale (34.3%) and Kiryandongo (14.1%) refugee settlements. Variations were reported for forest fires in the three settlements: Rhino Camp (85.2%), Kiryandongo (7.6%) and Nakivale (7.2%). Similar patterns of variation in the three settlements were reported for other climate-change-related hazards. (Table 3).

Table 3.

Climate change-related hazards experienced in the past 10 years, across the three refugee settlements

Variable	Category	Overall (N = 600)	Kiryandongo	Nakivale	Rhino
Hazards experienced in the past 10 years **	Drought	549	178 (26.8)	192 (47.5)	179 (36.4)
	Fire in the forest	186	14 (7.6)	8 (7.2)	164 (85.2)
	Community fires	194	38 (19.4)	13 (10.9)	143 (69.7)
	Flood	286	21 (5.6)	150 (65.3)	115 (29.1)
	Windstorm	285	60 (18.7)	67 (34.3)	158 (47.0)
	Thunderstorm	189	29 (14.1)	35 (27.9)	125 (58.0)
	Hailstorm	132	9 (5.2)	69 (65.2)	54 (29.6)
	Heavy rain	404	36 (7.1)	176 (56.9)	192 (36.1)
	Sporadic rain	187	31 (13.6)	69 (49.9)	87 (36.5)
	Erosion	358	76 (18.7)	87 (35.3)	195 (46.0)
	Landslide	143	6 (3.9)	25 (26.7)	112 (69.4)
	Heat wave	270	162 (54.4)	47 (26.0)	61 (19.6)
	Cold wave	150	25 (38.8)	32 (12.6)	93 (48.6)
	Median occurrence		14.1	34.3	46

**multiple choice response

Median occurrence – overall experience of at least one climate-related hazard in each settlement

Participants’ accounts complemented these quantitative findings, often describing experiencing increasingly unpredictable weather conditions. They often described these climatic changes as becoming more severe over time, with direct consequences observed and experienced for agricultural production, and the health and well-being of the community.

“[…] by then I used to see the rainy season was much closer, but nowadays sometimes it rains too early. The seasons are not predictable like before. If it has started raining in March, the other year it starts raining in February. Not like those years when I had arrived here in 2011, 2010. I guess the weather pattern has changed.” IDI-03-Rhino.

Climate change-related hazards, food security and related health outcomes

Climate change-related hazards were significantly associated with food insecurity and poor health outcomes among refugee households. The risk of food insecurity was higher among participants living in areas affected by forest fires (13%, aPR = 1.13; 95% CI = 1.08–1.23), unreliable rainfall (7%, aPR = 1.07; 95% CI = 1.01–1.15), and landslides (7%, aPR = 1.07; 95% CI = 1.01–1.14) (Table 4).

Table 4.

Climate change-related hazards associated with refugee food insecurity

Variable	Category	Food secure N = 92 n (%)	Food insecure N = 508 n (%)	CPR (95% CI)	p-value	aPR (95% CI)	p-value
Climate-related disasters in the past 5 years**	Drought	83 (15.1)	466 (84.9)	1.00 (0.94–1.07)	0.93	1.03 (0.94–1.13)	0.54
	Fire in the forest	4 (2.2)	182 (97.8)	1.11 (1.08–1.14)	p < 0.001	1.13 (1.08–1.23)*	< 0.001
	Flood	33 (11.5)	253 (88.5)	0.99 (0.96–1.03)	0.88	1.03 (0.98–1.09)	0.23
	Windstorm	37 (13.0)	248 (87.0)	1.05 (1.01–1.09)	0.013	0.98 (0.92–1.05)	0.64
	Thunderstorm	21 (11.1)	168 (88.9)	1.05 (1.02–1.09)	0.002	0.97 (0.91–1.04)	0.59
	Hailstorm	13 (9.8)	119 (90.2)	1.02 (0.98–1.06)	0.29	1.01 (0.95–1.08)	0.64
	Heavy rain	45 (11.1)	359 (88.9)	1.06 (1.00-1.12)	0.04	1.03 (0.96–1.11)	0.60
	Unreliable rain	20 (10.7)	167 (89.3)	1.02 (0.98–1.06)	0.26	1.07 (1.01–1.15)	0.03*
	Landslide	6 (4.2)	137 (95.8)	1.07 (1.04–1.10)	p < 0.001	1.07 (1.01–1.14)	0.02*
	Heat wave	54 (20.0)	216 (80.0)	1.01 (0.97–1.04)	0.60	0.97 (0.92–1.03)	0.24
	Cold wave	21 (14.0)	129 (86.0)	1.04 (1.01–1.08)	0.016	0.94 (0.88–1.01)	0.08
Food shortage due to climate-induced events over the past 10 years	No	52 (57.1)	39 (42.9)		1	1
	Yes***	40 (7.9)	469 (92.1)	1.40 (1.23–1.59)	p < 0.001	1.39 (1.22–1.56)*	< 0.001

**multiple choice question

*Statistically significant association

***Yes - if at least one disaster occurred

These statistical associations were corroborated by participant narratives of repeated climate shocks affecting food production, water access, and household wellbeing. Several participants described extreme temperatures, unreliable rainfall, and a significant decrease in rainfall over the years as disrupting planting cycles, reducing yields and increasing uncertainty around food production.

Extreme temperatures

Participants highlighted a connection between rising temperatures over extended periods and the occurrence of forest fires. They reported an overall increase in temperatures and a more frequent occurrence of prolonged heatwaves, particularly in 2020, which they considered a pivotal moment. These extended periods of high temperatures increased evaporation rates, resulting in crops drying out prematurely. The rise in temperatures signifies broader climate shifts that impacted not only agriculture but also the health and well-being of the community.

“Before that, rainfall was abundant, and sunshine was not excessive. But in 2020, we had a long period of sunshine whereby the temperature was high, and even the food we produced in the garden all dried up, and we could not get a good yield. That is in 2020. It affected the first season, that is when things were worse.” IDI-02-Kiryandongo.

It started from December 2020 to August 2023. It was a long dry spell and almost there was no production that time […] this greatly affected food availability at household level” KII-08-Nakivale.

Sporadic/unreliable rainfall

Participants similarly identified sporadic and unreliable rainfall as a key factor, noting that it has led to drought, floods, landslides, cold waves, and hailstorms. Several participants observed a significant decrease in rainfall over the years, with former patterns of two reliable rainy seasons now potentially reduced to just one. This reduction in rainfall has resulted in drought conditions that directly affect thermal discomfort, crop production, and water availability.

“So, in what we are seeing currently in Kiryandongo, the rainfall is reducing. We used to have two good seasons, and that was before we used to have two good seasons. Right now, the only reliable season for crop production is the second one, which starts in August. This first one is not always reliable.” KII-03-Kiryandongo.

Landslides and flooding

Participants also reported severe flooding and landslides due to extreme increases in precipitation and dramatic shifts between heavy rainfall and drought, with limited preparedness among communities in 2022 and 2023. These events impacted both crop production. The location of Rhino Camp near the River Nile and Nakivale in a basin made these areas more vulnerable to flooding from water accumulating in neighbouring high-altitude regions, while Kiryandongo communities, situated near water sources and those receiving more rainfall than others, experienced flooding.

“Remember that Rhino camp is located in an area around the Nile River. Yeah. So those things and the flood started happening in some of these areas, which we’re not so much, but recently we’ve had a lot of floods where crops and houses were destroyed in one of the areas in Rhino camp.” KII-07-Rhino.

“And one thing you also realise is that I think Nakivale is a cattle corridor. It is extremely dry, so when it rains, the surface runoffs are too much, so you find it will rain like in another location, and you’ll see the surface runoffs come in another location” KII-12-Nakivale.

While hailstorms did not show a significant association, several key informants noted that they occurred during the wet season, and while they were not a regular occurrence, they reportedly occurred more in Nakivale than in the Rhino and Kiryandongo settlements. Although they were not a regular annual occurrence, the damage they caused was perceived as localised to certain areas within the settlement and specific to certain plants, such as beans, maize, and bananas.

“During the wet season, there is a lot of rain, and sometimes the rain comes with hailstorms. It destroyed all their banana plantations. Actually, all the banana plantations were left destroyed, but maize and beans were the most affected.” KII-01-Rhino.

“Yes, 2023 actually, they had very varying patterns for like a week, it was extremely hot, then, boom, the rains came, all the crops on the ground in the settlements […] 0% production with hail storms actually” KII-12-Nakivale.

Pest and disease burden

Approximately half of the participants reported an increased incidence of vector-borne (n = 296; 49.3%) and water-borne (n = 303; 50.5%) diseases, respectively, among their family members over the past 10 years, while roughly one-third observed new diseases in crops (n = 191; 31.8%) and a rise in crop pests during the same period (n = 206; 34.3%). Additionally, general food shortages leading to undernutrition were reported (n = 509; 84.8%) (Table 5).

Table 5.

Reported increase in pests, food shortage, livestock, crop and human diseases over the past 10 years due to climate change and variability

Variable	Yes frequency (n = 600)	Percentage (%)
Increased livestock diseases	132	22.0
Increased incidence of any disease in your family	439	73.2
Increased incidence of vector-borne diseases in your family	296	49.3
Increased incidence of waterborne diseases in your family	303	50.5
New diseases in crops	191	31.8
Increased incidence of crop pests	206	34.3
Food shortage	509	84.8

Water-borne diseases

Participants described how the effects of climate change, such as water scarcity during droughts, prolonged dry spells, and lake flooding during the rainy seasons, posed significant health challenges for communities in various ways, including increased dehydration and waterborne diseases. Prolonged dry spells resulted in water scarcity, leading to dehydration. Furthermore, reduced water flow from boreholes led communities to explore alternatives such as relying on often-contaminated water sources like lake water or swamps. This led to an increased incidence of waterborne diseases such as cholera and diarrhoea.

“[…] when sunshine becomes much these boreholes be bringing little water and most of us fail to get water and children and other people lack blood it dries up in the body […] we go fetch from the lake but when sunshine becomes much even the lake dries up and water extends to where we cannot reach […] the water is always dirty and we get stomach pains and even infections” IDI-05-Nakivale.

“[…]There was an outbreak of cholera… the community was accessing water from the lake, and it was not suitable.” KII-05-Nakivale.

Vector-borne diseases

Consistent with the survey findings, which showed that nearly half of the refugee households reported increased vector-borne diseases, most participants narrated that the fluctuations between extreme heat and heavy rainfall created ideal conditions for mosquitoes, significantly increasing the malaria incidence. Moreover, they observed that stagnant waters resulting from frequent floods further exacerbated the situation by providing numerous breeding sites for malaria vectors.

“[…] That is very true because currently the malaria burden is high; the prevalence is standing at 54% in the settlement alone, so it is our biggest challenge in the settlement.” KII-05-Nakivale.

When there is too much rain like last season, issues […] and other infections such as malaria also come in” IDI-01-Rhino.

Undernutrition and deaths from food shortage

Survey findings showed that 84.8% of refugee households experienced food shortages, indicating widespread nutritional vulnerability across the settlements. As climate change intensified, the community’s ability to grow essential crops, a major source of vitamins and minerals, was also affected, ultimately worsening the refugees’ nutritional status. For instance, in Kiryandongo and Rhino, extreme sunshine and drought significantly hindered local communities’ ability to grow greens, ultimately impacting their availability and nutritional value.

“Children are malnourished […] like severe acute malnutrition and moderate acute malnutrition due to the lack of food, the lack of diet. […] Because the more you have to diet, the more you, I mean, you stay healthy.” IDI-07-Rhino.

“Eventually you realise they (refugees) are increasing in malnutrition programs because they don’t have enough in their consumption.” KII-08-Nakivale.

Furthermore, in extreme cases, starvation resulted in community deaths.

“Previously, we lost two family members because, according to the people who were our immediate neighbours, they said those people had set fire about two to three weeks ago. Meaning that they had not eaten food in three weeks that had passed. So, it resulted in death, and we recently buried them. So, the issue of food is really serious. […] very complex.” IDI-05-Kiryandongo.

Effects on food availability, accessibility, utilisation and stability

Table 5 above indicates that approximately 85% of refugee households reported experiencing food shortages. The qualitative study participants complemented the quantitative findings as expounded below.

Several participants reported experiencing periods of heavy rain that, over time, left crops waterlogged, resulting in stunted growth and poor yields, particularly in maize. Excessive moisture promoted the proliferation of pests and diseases, thereby reducing harvest quality.

“There was a period when there was too much rain, and the crop never did well because of too much water. Then, during the other period when there was a lot of sunshine, we didn’t get anything. People were starving because there was a shortage of food in the settlement.” IDI-02-Kiryandongo.

Furthermore, increased spoilage of yields post-harvest and improper weather conditions during the growth phase also resulted in crops that are less resilient to storage conditions, leading to higher spoilage rates as post-harvest processes were disrupted, thereby reducing the amount of food available.

“Yeah, it can be from the garden or in the storage, […] because in the garden is where food can get spoiled due to the weather change; this is not because the farmer wants it. A farmer might have pesticides and fertilisers, and even tries to weed at the right time, but the weather turns out to be bad, no rainfall, and too much sunshine.” IDI-01-Kiryandongo.

Other participants expressed concern about altered rainfall patterns and related high temperatures, which affected not only crop growth and yields but also the availability of staple foods such as maize and beans. They further noted that water scarcity made it difficult for greens and vegetables to grow, thereby rendering them unavailable in households and costly in the market. This affected dietary diversity among refugee households.

“Mostly for those who cultivate vegetables, when the drought comes, everything becomes dry. Accessing water, like going to fetch water for watering, you might fail the whole day; you don’t get water. And the more you stay without watering the greens, it dries.” IDI-03-Kiryandongo.

Discussion

This study, which assessed how climate change and variability-related hazards influence food security and health in three refugee settlements in Uganda, found that the occurrence of hazards such as forest fires, floods, and cold waves increased food insecurity. More than half of the participants reported an increase in the incidence of human diseases, and generally, these events were reported more frequently in Rhino and Nakivale than in Kiryandongo. Qualitative insights revealed three major themes: climate change-related hazards; pests and disease burden; and food availability, access, utilisation and stability. These findings highlight the heightened differential exposure and vulnerability of refugee communities to climate change-related food insecurity and adverse health outcomes. Mechanisms for the interventions could therefore be tailored to the different settings and environments of the refugee settlements to attain the best outcomes.

The finding of an association between forest fires and increased likelihood of food insecurity is consistent with existing studies [35, 36], which highlighted that forest fires can destroy agricultural land, crops and livestock, which directly impacts food availability, access and stability. Moreover, the long-term consequences of soil degradation, water contamination, and biodiversity loss due to forest fires can further compromise food systems [37]. Economic losses incurred due to fires can limit people’s purchasing power, exacerbating food insecurity [38]. Importantly, the increasing frequency and intensity of forest fires, linked to climate change and variability [39] could amplify these effects, creating a vicious cycle in which food insecurity contributes to further environmental degradation from economic activities such as deforestation for timber and charcoal production. Such a cycle heightens the vulnerability of refugee communities to disasters. This discovery underscores the need to develop effective early-warning strategies and methods to mitigate the drivers and impacts of forest fires on crop production and livestock, particularly in settlements like Rhino Camp that are exposed to extreme temperatures.

Flooding was another significant contributor to food insecurity in the refugee settlements. Communities located near major rivers were more susceptible to flooding and its associated negative impacts on food security. For instance, areas in Rhino Camp near the River Nile were particularly vulnerable to flooding due to riverbanks bursting, which led to soil erosion and crop destruction. Similarly, Nakivale’s position in a river basin makes the refugee communities and farmlands more exposed to flooding from water accumulation in neighbouring high-altitude regions, especially during seasons of unusually high precipitation. Flooding can also explain why Nakivale has the second-highest level of food insecurity despite being situated in Uganda’s most favourable climatic region (tropical), which is ideal for food production. Flooding exemplifies extreme weather events stemming from heavy or prolonged rainfall linked to climate variability [40]. These events result in reduced crop yields, diminished safe water availability, and instigate inflation in market prices, thereby heightening the risk of food insecurity within these communities. (Lucas et al., 2023).

Conversely, food insecurity was lower where cold waves were reported. These findings resonate with those from Tanzania, which revealed that food insecurity was less frequently reported in areas that experienced cooler temperatures, but contrast with evidence from Australia [41, 42]. Perhaps cooler temperatures reduce the prevalence of crop pests and diseases, resulting in improved agricultural yields. Moreover, cold waves may coincide with seasons of increased agricultural productivity, enhancing food availability. It is also plausible that these refugee communities may have had access to cold-tolerant crops or diversified food sources, which could have mitigated the effects of cold waves on food insecurity. In addition, since the definition of “cold waves” may differ across geographical regions, it is plausible that the extreme cold conditions as defined in the sub-Saharan African context may not be as deleterious to crop production as in other regions globally. In fact, cold weather is often synonymous with rainy seasons, favouring maximum crop production as agriculture is predominantly rain-dependent [43, 44]. These findings highlight the need to understand seasonal patterns when designing interventions against food insecurity.

Quantitative findings indicated a higher incidence of waterborne and vector-borne diseases, supported by qualitative evidence of increased illnesses such as malaria following climate change-related events. This aligns with a growing body of literature highlighting the intricate relationship between climate and health [45, 46]. Extreme weather events, such as floods, droughts, and heat waves, intensify conditions that are conducive to the spread of infectious diseases [47]. Higher temperatures can favour the growth of pathogens, while water scarcity caused by droughts impacts water quality and hygiene practices, particularly in arid and semi-arid regions, increasing the burden of diseases such as diarrhoea [48]. Furthermore, during floods, contamination of water supplies often leads to outbreaks, especially in areas with insufficient sanitation infrastructure. In addition, drought conditions can contribute to an increase in vector-borne diseases such as malaria [49]. While the relationship is plausible, our study did not assess such a significant association. Therefore, we cannot conclude with certainty whether the reported increase in these diseases is attributable to the effects of climate change. Nonetheless, the consistency of our findings with previous studies strengthens our evidence, underscoring the need for further research to examine the relationship between climate change and these health conditions in refugee settlements.

In the qualitative results, participants experienced an increase in malnutrition-related illnesses and deaths. Prolonged droughts and erratic rainfall patterns disrupt agricultural production, resulting in food shortages and, consequently, nutritional deficiencies [50, 51]. This is particularly concerning in refugee settlements, where food supply chains are already fragile [52]. Malnutrition compromises the immune system, making individuals more susceptible to infectious diseases [53] and poor maternal and child health [24, 54], which could explain the reported increase in disease incidence. Our results have significant implications for public health policy and practice, highlighting the need for investments in climate adaptation measures, such as drought-resistant crop varieties and community training in sustainable agricultural practices, to effectively mitigate food insecurity-related malnutrition.

Participants’ reports of altered seasonal cycles, characterised by prolonged dry periods and brief, intense rainy seasons, align with findings from Cameroon [55] and reflect the wider climatic changes affecting agriculture and food security across the three settlements. These shifts in climate patterns align with global trends, where climate change increasingly disrupts traditional agricultural practices and worsens food insecurity [56], especially in vulnerable populations, such as those in refugee settlements. Moreover, the reduction from two planting seasons to one, with the first season becoming increasingly unreliable, has several implications for food production and security. In agriculture-dependent countries such as Uganda, having multiple planting seasons is essential for ensuring a stable food supply throughout the year [57, 58]. Therefore, losing a single season may leave communities with fewer chances to recover from a poor harvest, increasing their vulnerability to food shortages. These findings underscore the need to invest in alternative livelihoods to sustain a food supply and reduce dependence on increasingly unreliable agricultural cycles.

This study provides new evidence regarding the effects of climate change and variability-related events on the most crucial aspects of refugee communities: food and health. Importantly, it illustrates refugee settlements with varying climatic, geographic, and socioeconomic characteristics, enriching our understanding and enhancing the generalizability of the findings across different refugee contexts. However, the study does have certain limitations. First, we employed a small sample size and thus our study was not well-powered to detect settlement-specific differences in food insecurity and related health outcomes. The post-hoc power tests indicated that our sample could only detect significant differences of approximately 15%. Our findings should therefore be interpreted as exploratory, underscoring studies with larger sample sizes. Secondly, the study relied on self-reported data, and participants were asked to recall climate change and variability-related incidents over a span of ten years, which may have introduced recall and possible social desirability bias. To minimise this, we employed a mixed methods design that triangulated multiple methods, data sources, and analysts. Third, the cross-sectional design used limited causal inference, as exposure and outcome variables were assessed simultaneously. Finally, the multilingual nature of refugee settlements may have affected data collection despite the recruitment and training of multilingual research assistants. In Rhino Camp specifically, some refugees predominantly communicate in Juba Arabic rather than Lugbara, which is more commonly spoken by host communities. The limited inclusion of Juba Arabic-speaking research assistants may have restricted participation among certain refugee households and introduced risks of miscommunication, information bias, and underrepresentation of some linguistic groups.

Conclusion and recommendations

Climate change and variability-related hazards such as forest fires, floods, and droughts may significantly threaten food availability, access, and stability, thereby heightening the vulnerability of refugee communities to food insecurity. Additionally, climate-related disruptions, including shifts in seasonal cycles and increased extreme weather events, may undermine agricultural productivity, worsen undernutrition, and increase the burden of vector-borne and waterborne diseases. The study findings suggest that the complex adverse impacts of climate change may negatively affect the well-being of refugees, particularly in settlements already grappling with poor structural and social determinants of challenges that influence health outcomes.

To strengthen the resilience of refugee settlements against climate-related hazards, it is essential to implement community-based early warning and disaster preparedness systems, particularly in high-risk areas such as Rhino Camp and Nakivale. These systems should be complemented by investments in climate-resilient agriculture, and training programmes in sustainable farming practices. Furthermore, seasonal adaptation strategies, such as aligning agricultural practices with meteorological forecasts and educating farmers about the implications of shifting seasonal cycles, are necessary to help communities respond to changing weather patterns. Upgrading water supply and sanitation infrastructure is equally critical to minimising health risks associated with floods and droughts, while the health system should scale up targeted nutrition and health interventions to tackle malnutrition and the increase in climate-sensitive diseases. Public health planning must incorporate climate risk assessments to effectively safeguard vulnerable populations. Lastly, further longitudinal research is needed to establish causal links between specific climate hazards and health outcomes in refugee settings. Such evidence will be vital for shaping effective policy and ensuring that interventions are both data-driven and context-specific.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HFIAS

Household food insecurity access scale

OPM

Office of the prime minister

UNHCR

United nations high commission for refugees

Author contributions

MTW conceptualized the study, performed formal analysis and wrote the first draft of the manuscript. JPG, JN, DM, AM, CM, MGBN, CGO, HB contributed to the methodology, study implementation and reviewed and edited the manuscript for substantial intellectual content. All authors read and approved the final version for submission.

Funding

The study was funded through an early-career grant from the Climate Adaptation Research Program (CARP), through the USAID Bureau for Humanitarian Assistance (Award# 720FDA20CA00006), and by The Kakande Ministries, under the Department of Philanthropy, Health and Welfare, to MTW.

Data availability

The dataset(s) supporting the conclusions of this article is(are) included within the article (and its additional file(s)).

Declarations

Ethics approval and consent to participate

This study was approved by the Makerere University School of Public Health Research and Ethics Committee (SPH-2023-488) and Uganda National Council for Science and Technology (HS3685ES). Permission to conduct the study within refugee settlements was obtained from the Office of the Prime Minister (OPM), Department of Refugees, Kampala. Prior to participation, all participants provided written informed consent. Participant confidentiality, privacy, safety, and anonymity were maintained. The study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.