NATURAL DISASTER A.

19 / MALAWI
Malawi 2015 / FLOODS
floods OVERVIEW
overview AFRICA

OVERVIEW MALAWI 2015 / FLOODS

CRISIS Malawi floods, January 2015.

KARONGA
UNITED REPUBLIC
TOTAL HOUSES
523,347 houses affected. 356,643 completely destroyed .
ZAMBIA
1 OF TANZANIA
DAMAGED
RUMPHI

1,101,364 individuals affected . 1

TOTAL PEOPLE
AFFECTED 336,000 individuals displaced (230,000 in displacement
sites; 106,000 in host sites)2.
MOZAMBIQUE
RESPONSE
LOCATIONS
15 districts affected (the most affected were Chikwawa, Nsanje and
Phalombe).
SALIMA

RESPONSE
50,000 households served with NFIs (70,000+ planned).
Approx.

Over19,000 households assisted with emergency shelter (32,000+ planned).

MANGOCHI
OUTPUTS
(as of August 2015) 3 NTCHEU
MACHINGA
Over 2,000 households assisted with repairs and retrofits (5,000+ planned).
BALAKA

BLANTYRE ZOMBA
CHIRADZULU PHALOMBE
1
Malawi 2015 Floods Post Disaster Needs Assessment Report, Gov. of Malawi, March 2015, http://bit.ly/2ogiYqQ.
MULANJE
2
UNDAC Assessment Report, 6 February 2015.
3
Data reported to the Shelter Cluster 4W by humanitarian organizations. Note that this data may not be 100% accu- CHIKWAWA THYOLO
rate nor complete (it does not include the figures of the overall response).
NSANJE
AFFECTED DISTRICTS

SUMMARY OF THE RESPONSE

The floods in Malawi in 2015 led to displacement and widespread damage to housing in the affected areas. Displacement sites
were set up in public buildings (such as schools) during the emergency phase, and assistance was provided primarily in these
sites. After the first few months, the focus shifted towards relocation and supporting return to IDPs’ places of origin, in order to
enable collective centres to go back to their functions, and facilitate early recovery. According to data reported to the Shelter
Cluster, emergency shelter support consisted mainly of distribution of tents and tarpaulins, while repairs assistance was primar-
ily in the form of tool kits and/or materials, coupled with trainings.

9 JAN 2015
2016
TIM E L IN E

1 2 3 4

FLOODS EMERGENCY RELIEF

EARLY RECOVERY RECOVERY

RECOVERY

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

1 13 Jan 2015: Declaration of Sate of Disaster by the Government of 3 2 Mar 2015: Rapid joint assessment released by Shelter Cluster.
Malawi.
2 22 Jan 2015: Shelter Cluster released Preliminary Response Plan. 4 End of 2015: Deactivation of Malawi Shelter Cluster.

CONTEXT AND BACKGROUND clared on 13 January 2015. As a result of the prolonged,

80% of the population of Malawi live in rural areas. The heavy, rainfall, the Shire River reached its highest level in 30
economy is primarily agricultural, which accounts for 90% of years, bursting it banks in multiple areas.
export revenues. National GDP per capita is one of the low-
est globally and the economy has experienced low growth. SITUATION AFTER THE FLOODS
Malawi is also heavily reliant on investments from global fi- The extreme rainfall event and resulting flooding led to dis-
nance institutions. A lack of trust in the Malawian Govern- placement, with many affected households evacuated to col-
ment by these institutions (since 2013) has led to a reduction lective centres (schools, churches and mosques). As these
in investments, further stagnating economic growth. naturally (and in some cases enforcedly) disbanded after the
first two months, affected households with no long-term shel-
Malawi experienced above-average rainfall throughout De-
ter solution constructed simple emergency shelters, or stayed
cember 2014 and January 2015. The Southern Region of
with host families.
Malawi received 400% more rainfall than the Long Term
Mean for the region. 15 of the country’s 28 districts expe- Properties sustained damage through a combination of rain
rienced significant flooding, with a state of emergency de- and high winds. The most affected communities were more

SHELTER PROJECTS 2015 - 2016 79

 AFRICA A.19 / Malawi 2015 / floods overview NATURAL DISASTER

© Julien Lefevre
After several weeks of heavy rains, the Shire River reached the highest level in the past 30 years, burst its banks in several locations and caused widespread flooding.

vulnerable to the disaster, as a result of their shelter and settle- The Shelter Cluster’s initial strategic objective was to relo-
ment typologies. Many of the inhabitants of the flooded rural ar- cate all people from collective centres into planned camps
eas resided in one-storey houses, constructed using traditional or resettlement areas5. Expected outputs and impacts of the
techniques and materials, such as sun-baked mud-bricks and emergency response were:
thatched roofs. The flooding, rainfall and wind caused homes
• 31,636 households provided with tents and NFIs.
to disintegrate and roofs to blow off. There appeared to be a
correlation between the degree of damage sustained and the • Assessments conducted in all the 15 districts for strate-
construction techniques used. As shown by the Rapid Joint As- gic positioning of camp sites.
sessment (March 2015), 47% of houses built with fired bricks • Displaced people in the camp sites to be trained in con-
and CGI roofs reported damage, compared to 71% of those struction, for dignity and Disaster Risk Reduction.
built with sun-baked bricks, and 78% of wood and mud houses. • Resettlement areas properly laid out.

EMERGENCY SHELTER PHASE EARLY RECOVERY PHASE

The Shelter Cluster, led by the Ministry of Lands, Housing and By early March, the government prioritized the closure of
Development, was activated shortly after the emergency, and camps and the return of IDPs. This change in approach led
a Rapid Joint Damage Assessment was undertaken by vari- to a swift re-focusing from emergency operations to early re-
ous clusters4. covery planning within the humanitarian community. As part of
The international organization co-leading the Cluster quickly these efforts, the Shelter Cluster led the process of preparing
established a large shelter pipeline, and the first significant a “Durable Solutions Framework” and, where feasible, orien-
shelter distributions took place in early February, with tents tated its own efforts towards providing shelter support in are-
and shelter kits being airlifted to areas on the east bank of as of return. Supporting the ability to return was essential
the Shire River that had been completely cut off by the floods. to encourage livelihood recovery and to allow collective
centres to return to their proper use. The Cluster aimed to
During the emergency phase, the government promptly erect- provide adequate shelter in the camps, whilst also strengthen-
ed tents in the most critical displacement sites, in order to clear ing the capacity of the displaced population for early recovery,
the public facilities, particularly schools. The sites were selected through training on good construction methods and through
without sufficient planning and the tents set up hurriedly, lead- the provision of construction materials.
ing to challenges such as overcrowding and gaps in WASH and
Protection. Additionally, the distribution of humanitarian aid was The Cluster and the government promoted the use of fired
reported to create a draw to these sites, partially driven by the bricks (as opposed to sun-dried bricks) for reconstruction, so
underlying poverty and also by the food insecurity, created by that buildings would be more resistant to disintegration6. How-
flood damage to crops and livelihoods. A further challenge in the ever, a lack of availability of wood to fire the bricks (or financial
response was that initial assessments and distributions tended resources to purchase fired bricks) led to many households
to neglect IDPs in host communities, which increased the draw resorting to unsafe traditional building approaches. Some
to displacement sites and complicated coordination efforts. households received shelter assistance from government
and NGOs in the form of shelter kits (tools and tarpau-
4
The Assessment is available at http://bit.ly/2jbPHqw lins), tents, or materials to construct temporary timber
and plastic-sheet shelters. In assessments conducted by
humanitarian organizations, communities expressed a prefer-
ence for basic materials and tools, to repair or construct core
dwellings supplemented by local materials, including earth
blocks and grass thatching. This was considered an appro-
priate and durable solution to their immediate and longer-term
shelter needs, which would also allow them to focus on their
priorities, i.e. food security and livelihood recovery.
© Jean-Philippe Antolin

The case studies that follow show two approaches taken by

humanitarian organizations. While the first (A.20) was a short-
term project focused on the emergency relief and early re-
covery phase, the second (A.21) was a longer-term recovery
programme looking at housing reconstruction, with significant
Displacement site at Bitilinyu. These collective centres were the initial option for training and Disaster Risk Reduction components.
those who had to leave their homes and caused a significant draw, due to the
distributions of aid (and relative neglect of IDPs in host settings). These sites 5
Preliminary Response Plan, released on 22 January 2015 (http://bit.ly/2i0oiKI).
were also particularly overcrowded and had gaps in protection and hygiene. 6
Key Shelter Safety Messages - 2015 Malawi Floods and Storms.

80 www.shelterprojects.org SHELTER PROJECTS 2015 - 2016

 NATURAL DISASTER A.20 / MALAWI
malawi 2015 / FLOODS
floods AFRICA

CASE STUDY MALAWI 2015 / FLOODS

KEYWORDS: Emergency shelter, NFI distribution, Early recovery, GBV risk mitigation

CRISIS Malawi Floods, January 2015

UNITED REPUBLIC
TOTAL HOUSES
DAMAGED 523,347 houses (Source: Gov. of Malawi). OF TANZANIA

230,000 in displacement sites

TOTAL PEOPLE
DISPLACED 106,000 displaced in host sites
ZAMBIA
(UNDAC assessment report, 6 February 2015).

PROJECT LOCATIONS Chikwawa, Zomba, Mulanje.

BENEFICIARIES 1,874 households. MOZAMBIQUE

1,224 tents with household NFI kits

650 shelter kits
PROJECT OUTPUTS 500 tarpaulins
960 solar powered lamps ZOMBA

20 packs of classroom materials MULANJE

CHIKWAWA
100% of shelters distributed were verified as received.
OUTCOME INDICATORS 67% of respondents living at their (or a new) home site at
the time of ex post evaluation (Oct 2015), compared to 4%
at time of distribution.
PROJECT AREAS

3.6 m2 per person (based on national

SHELTER SIZE 1 Tent = 18.5m2 SHELTER DENSITY
average household size of 5.1).

MATERIALS COST USD 313 per household. PROJECT COST USD 550 per household.

PROJECT SUMMARY

This project had a relief-oriented and a recovery-oriented outcome objective. Through the provision of tents and shel-
ter-related NFIs, it aimed to meet immediate shelter needs and enabled affected households to move out of gender-seg-
regated collective centres, supporting return and easing overcrowding. In order to support early recovery, tarpaulins and
fixing kits were distributed to build or repair shelters, coupled with basic training and tools to assist with reconstruction
or earning a livelihood.

9 JAN 2015 NOV 2015

1 2 3
TI M E LI N E

PLANNING EMERGENCY RELIEF DISTRIBUTIONS

FLOODS RELIEF AND EARLY RECOVERY PHASE EX POST EVALUATION

JAN FEB MAR APR MAY OCT

1 22 Jan 2015: Shelter Cluster released Preliminary Response Plan. 23 Apr 2015: Distributions of tents to new caseload with totally de-
3
stroyed home completed. Distributions of shelter kits and tarpaulins to
2 22 Feb 2015: Distributions of tents completed. households with partially destroyed homes completed.

STRENGTHS WEAKNESSES
+ Reduced issues and risks related to overcrowding in collective - The recovery capacity of affected households was not properly
centres. understood.
+ Facilitated the return to areas of origin / own plots. - Lack of appropriate technical training to some recipients of the kits.
+ Responded at scale with different modalities. - Tarpaulins distributions did not include fixing kits.
+ Supported early recovery. - Detailed Post-Distribution Monitoring was not undertaken after the
relief distribution.

SHELTER PROJECTS 2015 - 2016 81

 AFRICA A.20 / malawi 2015 / floods NATURAL DISASTER

© Phil Duloy
Tents were used to clear the overcrowded collective centres during the emergency relief phase. By doing so, this project managed to address some of the issues faced
by the displaced people in those sites, including family reunification and reduction of hygiene and protection concerns.

CONTEXT
For more background information, see overview A.19.
Extreme rainfall in Malawi during January 2015 caused wide-
spread displacement, forcing households to seek immediate
shelter in collective centres. In rural areas, the flooding also
led to the destruction of harvests and damage to water sourc-
es, further exacerbating food-security issues. This created an
additional draw to collective centres. Displacement sites be-
came crowded, with a lack of basic services, such as water,

© Derek Locke
sanitation and hygiene, prompting concerns about the out-
break of diseases. The shelter sector was urged to respond
in a way that provided immediate lifesaving shelter (alongside
appropriate services) and increased the affected community’s Some structures sustained significant damage due to the floods.
capacity for early recovery.

RELIEF PHASE EARLY RECOVERY PHASE

During the initial phase of this intervention, the organization re- After the initial emergency phase, the project shifted em-
sponded to the immediate shelter needs at collective cen- phasis towards supporting early recovery. In Zomba and
tres. Due to severe overcrowding, there were concerns about Mulanje, shelter kits or tarpaulins were distributed to house-
Gender-Based Violence (GBV) and child protection issues, as holds with a partially destroyed home. Beneficiary house-
well as health issues resulting from a lack of basic services. holds were able to use tarpaulins and fixing kits to repair
People were living in gender- and age-segregated rooms, and and weatherproof shelters, until access to resources allowed
in some instances men were required to sleep outside. Tents them to seek a more durable solution. As part of the distri-
and NFI kits, all imported over several rounds, were distributed bution, a basic level of training was provided on how to use
to households verified as having a totally destroyed home. the items to improve structures. Repairs included fixing dam-
aged external walls and replacing roofs that had blown off.
The organization aimed to support households as part of
a return scheme, motivated by the government’s desire to In Mulanje, during the later stages, households with totally de-
decongest overcrowded collective centres. For those house- stroyed homes, that were still in collective centres or with host
holds who did not want to return to their previous site due to families, were also provided with a shelter kit. In such instanc-
flood risks, the team worked with the local Traditional Author- es, households received lumber and made basic shelters on
ity, the District Government and beneficiaries to identify safer their home site. During the ex post evaluation, all interviewed
areas of land. In some cases, most notably in the district of families had completed – or were in the process of improving
Zomba, water inundation prevented households from re- – their structures. Many of these households also reported
turning home. In such situations, tents were distributed and they wished to reuse the tarpaulins as roof of the new shelter.
implemented in spaces surrounding the collective centres.
Due to land restrictions, the number of tents that could be dis- LOCATIONS AND BENEFICIARY SELECTION
tributed was limited, when compared to the caseload at the The organization focused its efforts in more remote regions and
centres. In such cases, the organization identified beneficiary rural communities, where fewer humanitarian actors were op-
families based on agreed vulnerability criteria. As the levels erating and gaps in the response were soon identified. Com-
of rainfall dropped and waters receded, distribution teams munities were selected in coordination with the government-led
worked with beneficiary households so that tents could be re- Cluster. The district government identified the worst-affected
located and families could return to their home sites. communities that had not yet been reached by other actors,

82 SHELTER PROJECTS 2015 - 2016

NATURAL DISASTER A.20 / malawi 2015 / floods AFRICA

© Varshana Trudgian
© Malcolm Shead

By pitching the tent on their home site, people were able to start to rebuild their Tarpaulins from the shelter kits were used, amongst other purposes, to seal off
damaged houses. Supporting return was essential for enabling early recovery. damaged parts of the houses.

making assessments available to field teams. Assessment and understand the selection criteria, and felt that targeting was
distribution teams would then work with a local representative political in nature.
for the community, often a camp coordinator. The camp com-
MAIN CHALLENGES
mittees, appointed by the district government, would generate
a beneficiary list based on agreed criteria: totally or partially de- The distribution of humanitarian aid created a significant pull
stroyed home and, in some situations, additional vulnerabilities. factor towards collective centres. Identifying the benefi-
The organization’s assessment team cross-referenced the lists ciaries who genuinely required shelter assistance – from
with data compiled by the regional government and also under- those who were trying to access other items – proved problem-
took key informant interviews, to verify that the criteria had been atic. Flooding in Malawi had washed away crops, exacerbat-
applied appropriately and to mitigate selection bias. ing underlying conditions of poverty, and since food and other
items were being distributed at collective centres, it was felt that
PROJECT IMPLEMENTATION TEAM some households had registered in order to qualify for food aid.
This project was managed a by a full-time project manager This exacerbated problems associated with severe over-
based in Blantyre, with coordination and strategic responsibili- crowding. Sanitation was insufficient, families were forced
ties. This role was filled by a series of overseas staff posted for to split, and there were incidences of skin and other commu-
around four weeks at a time. Two sub-teams (each comprising nicable diseases. The urgency of lifesaving assistance was
four staff and volunteers from the organization’s roster) were stressed in the preliminary response plan, along with the deci-
located in the target districts, to manage the implementation sion to encourage return by supporting families at their home
and coordinate with the district government and other actors sites, which helped to reduce the draw to collective centres.
operating in the same region. Overall, 40 overseas staff and
Due to underlying resource deficiency and also the scale of
volunteers were involved in the response. At the field level,
the crisis, there was a lack of access to sufficient lumber
teams used a high number of local staff and volunteers to
in the emergency phase, for shelter kits to be easily deployed
assist with the implementation. Some of these were drawn
to a large percentage of the affected population. The deploy-
from other organizations, while others were recruited direct-
ment of tents enabled rapid distribution, allowing the im-
ly from the affected communities, and worked as translators
mediate easing of collective centres. Households could erect
and enumerators, assisted with distributions, training and tent
them on their land quickly. Where water inundation prevented
erection. In some instances, agreements were formalized
return to home, tents could be erected temporarily on land
through the creation of MoUs with the appropriate organiza-
adjacent to the collective centres. As the ground began to dry
tion. However, in situations where this did not happen, the lack
out, tents were moved to beneficiaries’ home sites.
of signed documentation caused issues during the implemen-
tation. For instance, newly posted staff or volunteers were A significant number of families who did not qualify for tents
not always clear on the agreed per-diem rates for distribution (according to the beneficiary criteria) had a severely damaged
teams. Consequently, the organization became stricter in the house, therefore being exposed to rainfall and high daytime
formalization of working relationships. temperatures. These households remained without adequate
shelter, as many did not have the resources to make simple
ENGAGEMENT OF AFFECTED PEOPLE improvements and repairs in a timely fashion. This influenced
Distribution teams from the local community were trained in the decision to distribute shelter kits alongside basic training
the erection of tents and were tasked with assisting benefi- in the second phase, and helped to reduce the issues of ineq-
ciary households. These teams also assisted with the reloca- uity felt by those who had not received any assistance.
tion of tents from collective centres to households’ home sites.
Although the organization coordinated well with the camp WIDER IMPACTS OF THE PROJECT
committees, more efforts should have been made to work A government representative commented that by promoting
more closely with the wider affected communities, particularly return to home – and distributing at people’s home sites or
in terms of communication and sensitization with non-ben- assisting with relocation – this project allowed to clear a
eficiary groups. Several cases were uncovered, during the number of the collective centres and their timely return
ex post evaluation, where community members had not ful- to their normal uses. This had a positive impact on the wid-
ly understood the organization’s goals and mission. In these er relief effort, beyond the shelter sector, and supported the
instances, families who did not receive assistance did not early recovery of communities following the flood events.

SHELTER PROJECTS 2015 - 2016 83

 AFRICA A.20 / malawi 2015 / floods NATURAL DISASTER

STRENGTHS, WEAKNESSES AND LESSONS LEARNED

STRENGTHS - Adequate technical training on the use of the shelter kit

was not always provided to beneficiaries. This was due, in
+ The project provided a mechanism for rapidly reducing
part, to the general lack of understanding (by implementing
the problems associated with overcrowding at collective
teams) of techniques associated with the shelter kit. Following
centres, with the distribution of tents and selected NFIs. It did
the completion of this project, shelter kit trainings were rolled
so by 1) Reducing exposure to vector and water transmit-
out across the organization’s network of staff and volunteers.
ted diseases; 2) Improving privacy; 3) In many cases facili-
tating the return home and therefore reinstating livelihoods - Tarpaulins were not distributed with a fixing kit, except
and supporting early recovery; 4) Mitigating risks associ- when part of the standardized shelter kits. Although there
ated with GBV and child protection, as well as enabling the were many cases were beneficiaries were still able to use
reformation of the family unit, with parents better able to these items to good effect, in some instances tarpaulins were
watch over minors. Qualitatively, beneficiaries reported this used for non-shelter purposes – such as drying food. This is-
to be an important outcome of the intervention, as certain sue may have arisen because beneficiaries did not receive
aspects of normal household behaviours could commence. the fixings required to utilize tarpaulins as intended, or due to
+ The organization was able to reach a greater number of a lack of training.
households and reduced the potential for inequity resulting - The early emergency phase did not include detailed
from the distribution of tents alone, thanks to the distribution post-distribution monitoring. This further affected the or-
of shelter kits or tarpaulins for those with a partially dam- ganization’s understanding of the barriers to early recovery.
aged house. This also yielded further positive outcomes
in terms of supporting early recovery.
MATERIALS LIST
Quantity Total Unit Cost
WEAKNESSES Materials
(per HH) Quantity (USD)
- Vulnerability assessments did not inform an under- Family Tent 1 1,224 276.9
standing of the self-recovery capacity. Early elements of Shelter kit (IFRC
1 650 51.5
the response were focused almost solely on immediate relief, specification)
and failed to consider the longer-term recovery needs of the af- Tarpaulin (IFRC
1 500 11.8
fected population. Whilst some beneficiaries were able to use specification)
the provision of emergency shelter as a platform for recovery, Household water filtration 1 500 32.3
others were unable to transition towards a more durable shel- Solar light 2 3,408 9.5
ter within the life cycle of the tent. As tents cannot be easily Blanket 5 6,120 7.3
adapted, this raised concerns that some beneficiaries would Kitchen set 1 1,224 23.2
become exposed to shelter-related issues at a later date. Mosquito nets 2 2,448 4.4
Jerry can (10l) 2 2,448 3.5
Tool kit 1 1,224 14.1

LEARNINGS

• Limited availability of food, inflated prices and a reduction in livelihood activities had a significant impact on the ear-
ly recovery capacity of affected households. If access to food and livelihoods is a known issue, this should be recognized
and included in assessments. In this response, evidence suggested that many households were drawn to collective centres
as a result of damage to crops, thus the linkages between shelter need and food poverty could be assumed from the outset.
• Vulnerability and capacity assessments should include pre-disaster secondary data, as well as post-disaster
secondary and primary data, and this should be factored into any resulting project design. Providing a household with
emergency shelter and NFIs can often provide the appropriate platform to begin the process of self-recovery. How-
ever, there are contexts when the pre-disaster conditions significantly inhibit the ability of the affected communities
to engage in self-recovery. Early, vulnerability-driven, emergency shelter, distributions need to be followed
by further capacity assessments and, if appropriate, an additional recovery-oriented component1. Although this
intervention provided immediate support for those at greatest risk as a result of the displacement, there should have
been more recognition of the impact of vulnerabilities on the capacity of households to recover.
• Detailed post-distribution monitoring should be undertaken to recognize specific vulnerabilities early on, and
enable the organization to provide an additional level of assistance, or link the most vulnerable beneficiaries with other
shelter actors. Although some informal checks were undertaken in the days following distributions, these were carried
out with the aim to identify any immediate gaps in provision, or to address aid-related issues. However, the evidence
gathered during an ex post evaluation showed that, due to underlying conditions of poverty, many households lacked
the material, financial or physical resources to transition towards a more durable form of shelter.
• Shelter kit interventions that do not include the appropriate level of technical training have a significantly lower
chance of yielding positive shelter related outcomes (both short- and long-term).
1
This approach was taken in project A.40 in response to the Ecuador earthquake.

84 www.shelterprojects.org SHELTER PROJECTS 2015 - 2016

 NATURAL DISASTER A.21 / MALAWI
malawi 2015-2016 / FLOODS
floods AFRICA

CASE STUDY MALAWI 2015-2016 / FLOODS

KEYWORDS: Core housing, Housing repairs, NFI distribution, Training, Guidelines, Disaster Risk Reduction

CRISIS Malawi Floods, January 2015

UNITED REPUBLIC

523,347 houses (Source: Gov. of Malawi).

TANZANIA
OF TANZANIA
TOTAL HOUSES DAMAGED

TOTAL PEOPLE AFFECTED 1,101,364 people (Source: Gov. of Malawi).

TOTAL PEOPLE DISPLACED 336,000 people (UNDAC assessment report). ZAMBIA

PROJECT LOCATIONS Zomba, Phalombe and Machinga districts.

BENEFICIARIES 1,090 households.

MOZAMBIQUE

1,090 houses benefited from emergency re-

pair and reconstruction activities.

9 model homes built in different communities,

to be replicated.
MACHINGA
109 cash-for-work grants provided to vulnera- ZOMBA
PROJECT OUTPUTS ble households (10%). PHALOMBE
Training provided to local builders and staff (in-
cluding 66 building supervisors, and three Trainings
of Trainers with 30 builders and 8 programme sup-
port staff).
Development of a training curriculum for
builders. PROJECT AREAS

SHELTER SIZE
22m and 16.5m (Size of house dependent on fam-
2 2

ily size, assumed at 3.5m2 per person). PROJECT SUMMARY

SHELTER DENSITY 3.5m2 per person. The programme aimed to assist flood-af-
fected people to return to their homes,
MATERIALS COST USD 200 per household. through the repair and reconstruction of
houses. This was done through the supply
PROJECT COST USD 552 (inclusive of training and dissemination). of tools, materials and technical training. It
also included training and information shar-
The majority of affected households returned to the ing to the community on more durable and
OUTCOME INDICATORS
site of their original dwellings, where possible.
resilient housing-construction methods.

25 JAN
9 JAN 2015 2016 2016

1 2 3 4 5
T I ME LI N E

FLOODS PLANNING IMPLEMENTATION HANDOVER

FEB MAR APR MAY JUN JUL AUG NOV JAN

1 Jun 2015: Initial discussions with community 3 Aug 2015: Training roll-out

Jul 2015: Beneficiary selection, shelter workshops with local build- 4 Aug 2015: Distribution of tools and materials
2
ers, development of shelter designs and training curriculum
5 Sep 2015 onwards: Continued technical support for reconstruction
STRENGTHS
+ Increased technical skills of local communities in the construction
of durable houses.
+ The programme provided an affordable housing solution. WEAKNESSES
+ Resources were used directly for housing recovery, accelerating - The project did not cater for all income levels.
the overall process of recovery. - Lack of organizational experience in shelter projects.
+ Model houses provided a reference for locals to replicate. - Poor planning led to delays in beneficiary selection.
+ The programme recognized traditional skills and knowledge. - Lack of adequate market assessment led to procurement challenges.
+ Capacity-building of local partners.

SHELTER PROJECTS 2015 - 2016 85

 AFRICA A.21 / malawi 2015-2016 / floods NATURAL DISASTER

© Jamie Richardson
People were given technical trainings and built model houses as part of the project. Here they are working on setting out the foundations of a model house..

CONTEXT
See overview A.19 for background information.

LOCATIONS AND BENEFICIARY SELECTION

The organization selected the three target districts due to the
high level of damage and the continued flood risk. Additionally,
the local partner had a strong presence in these districts and

© Jamie Richardson
good relations with the communities.
Priority was given to areas at greatest risk of future flooding
(confirmed by flood risk data), where most houses were dam-
aged or destroyed, and that had substantial loss of crops and
livelihood and fewest alternative income opportunities.
Workshops were held in target communities to identify best practices and devel-
Household selection was carried out in partnership with the op contextualized training for the community.
government District Offices and Traditional Authorities and fur-
ther verified by household visits. Priority was given to the most develop a training curriculum for other builders and to share this
vulnerable households, based on criteria including: single- and with their whole community. During the week, a complete core
child-headed households, elderly, disabled, households affected house was constructed, along with the provision of a curriculum
by HIV and low-income families with children under 5 years. and supporting communication materials.
The project aimed to advance gender equality and female The builders were then engaged to construct houses for
empowerment against cultural discriminatory norms, involving the most vulnerable families in each of their communities,
women in masonry and building workshops. which also provided a further training opportunity and model for
demonstration. Partner field staff and the builders also provided
PROJECT IMPLEMENTATION technical support to families during the construction.
Overall, the project was implemented with 52 staff members
and builders from a local partner which undertook work at the COMMUNITY PARTICIPATION
community level, while the organization provided a total of sev- In order to build upon existing knowledge and practice, the or-
en national and international staff for logistical support, coordi- ganization worked in partnership with communities and local
nation and overall supervision. builders, who were engaged from the outset in helping to re-
fine the affordability of the programme and then share their lo-
An initial shelter and housing assessment was undertak-
cal knowledge on construction practices and building materials.
en, highlighting that a number of proposed house designs were
not affordable and, if adopted, would only support a limited Throughout the programme, the organization maintained this
number of families. Given the prevalence of flooding and the collaboration through local and traditional authorities, focus
need to maximize the scale of the project with the available groups, workshops and household-level support.
funds, the organization aimed at supporting families to recon-
struct their permanent dwellings, using low-cost, locally availa- COORDINATION
ble materials, supplemented with in-kind assistance. Technical The project worked closely with the Shelter Cluster to agree on
training and support were also provided to identify and build the areas where the organization and its partners could work,
upon existing best local building practices, and to share this and to ensure that the approach was in line with Cluster proce-
information with the whole community. In order to do so, a se- dures. The Cluster Coordinator attended training sessions and
ries of workshops were held at central locations in the target assisted in parts of the training programme. All the materials
communities. Two builders from each community in the area developed during the programme were shared with the Cluster.
joined the workshop along with women and local government District government and traditional authorities were involved in
staff. The workshop included theory, discussion, site visits and identifying the communities, and communication was carried
practical exercises, to identify best construction practices. At out through them. The communities were then actively involved
the end of each day, the learning was recorded and used to in deciding the approach for the project.

86 SHELTER PROJECTS 2015 - 2016

NATURAL DISASTER A.21 / malawi 2015-2016 / floods AFRICA

© Jamie Richardson
© Jamie Richardson

Model homes were built, according to traditional designs. Additionally, materials Many of the traditional houses withstood the floods, as they were built with con-
were provided and cash-for-work grants for the 10% most vulnerable households. textually appropriate features, such as roof overhangs and raised platforms.

The project provided technical solutions, including refine-

ments to traditional house design, so that the roof could con-
tinue to be supported by the veranda posts, should the earth
walls collapse. During the workshops, emphasis was given
to soil selection for making adobe bricks and the correct
brick-making processes. The reason why many buildings
collapsed was due to the insufficient thickness of the walls,
therefore the improved design increased this width (from 10
to 15cm) so that the walls were more stable. It also ensured
© Jamie Richardson

that internal walls had proper foundations and were connected

to the outside walls, to further strengthen the structure.

DISASTER RISK REDUCTION

The communities were prone to heavy rains, high winds and
Trainings included the identification of good soils and mixing for block making. flooding. Therefore, Disaster Risk Reduction was very strong-
ly embedded throughout the programme. Community saf-
er-building information was disseminated to educate, inform
MATERIALS and provide examples. Other strategies were also encour-
All materials were purchased from within Malawi, largely aged, including planting trees to protect against driving high
through local markets. Timber supplies came from other dis- winds and rain. Trees could also be used as building materi-
tricts where trees were available for construction use, so as als or for firewood. Information was provided on Safer Earth
not to damage the local environment. Building for Floods and Rains, as a simple booklet and train-
ing curriculum for builders. This included information on haz-
Materials such as burnt bricks, cement and corrugated iron
ards, appropriate site selection and construction techniques
sheet roofing were beyond the financial means of the poorest
to reduce flooding in houses, as well as appropriate protection
households. Therefore, for wider impact, assistance had to be
and maintenance of houses and the environment.
focused on building solutions using local materials that were
affordable, replicable and achievable by the most vulnerable
MAIN CHALLENGES
and at-risk households.
The organization needed to convince government person-
While earth for brick-making and grass for thatching were lo- nel, politicians and other organizations that houses con-
cally available, other materials and tools had to be purchased. structed from local materials could provide a sufficiently
Communities were offered restricted cash to purchase ma- durable solution. This challenge was overcome mainly by
terials that were not freely available, but there was an over- building model houses that demonstrated this potential.
whelming request for in-kind support due to the distances to
markets, the capacity of markets, the cost of transport and the Additionally, extra technical support was brought in during
needs of families to focus on agricultural activities. the implementation process, to strengthen the local part-
ner’s capacity.
HOUSING DESIGN AND TECHNIQUES
WIDER IMPACTS OF THE PROJECT
Many houses had survived with little or no damage, even after
weeks of standing water, including those constructed using The programme explored and built upon existing local
earth brick and render. This is because these traditional hous- knowledge and practices, which enhanced the ownership
es had raised platforms that protected the core structure from and commitment of the residents and ensured that any rec-
erosion, and the veranda and large roof overhangs ensured ommendations were site-appropriate. The resources and
that the gables and walls were protected. This design, devel- information produced were shared with the Shelter Cluster,
oped over centuries, provided protection from the elements so that other actors could use them. Ultimately, this ap-
and, other than some minor repairs to the veranda and walls, proach provided a practical, inexpensive and replicable
allowed many families to return to their homes once the floods model to respond to similar flood events, in this and other
had subsided. parts of the country.

SHELTER PROJECTS 2015 - 2016 87

 AFRICA A.21 / malawi 2015-2016 / floods NATURAL DISASTER

STRENGTHS, WEAKNESSES AND LESSONS LEARNED

STRENGTHS
+ Increased technical skills of local builders in construc-
tion of durable houses, thanks to workshops conducted at the
community level.
+ The programme allowed for a more durable emergen-
cy response, using an affordable solution that would help
withstand future flood risks, yet was accessible by the poorest
and most vulnerable households.
+ Resources were used directly to support housing re-
construction, accelerating the overall recovery process,
instead of providing emergency or transitional support first.
+ Model houses provided a reference for locals to replicate.

©Jamie Richardson
Communities have started building houses using the safer
building guiding principles based on the model houses, which
therefore had a wider impact by providing a reference for other
members of the community.
+ The programme recognized traditional skills and knowl-
edge as an affordable and effective means of coping with Local materials were provided, as listed in the BoQ below.

heavy rains and floods, managing to convince locals that

these traditional methods were a good alternative to more
expensive materials, such as burnt bricks or concrete blocks. LEARNINGS

+ Increased capacity of the local partner. • Visible sections of the programme distract from
wider goals. The hard components of the pro-
gramme, such as the distribution of materials and
the construction of model houses, have the poten-
WEAKNESSES tial to dominate the programme and divert from the
- The programme did not cater for all income levels, as it wider objective of supporting the whole community
only provided a low-cost solution and did not consider those (by encouraging safer building practices and sup-
who could have afforded more durable housing. plying relevant information).

- Lack of experience in shelter projects of the organiza- • Importance of strengthening the organization’s
tion’s country programme and local partners meant that this capacity in varying sectors. The organization’s
had to be developed during implementation. preparedness needed to be reviewed to better re-
spond to future disasters, particularly with regards
- Delays in beneficiary selection and verification process to technical support, number of staff, as well as in
caused by poor planning slowed down the implementation. conducting beneficiary surveys to be used during
the identification and selection processes.
- Lack of adequate market assessment. There were some
logistical challenges in finding doors and windows, as no • Multisectoral programming, beyond shelter.
large supplier could be found. The programme should have also covered aspects
such as restarting livelihoods and food security, to
address family needs of those who were keen to
return home earlier than others.

MATERIALS LIST PER MODEL HOUSE

Ref Details Unit Quantity Unit cost (MK) Unit cost (USD) Total cost (USD)
1 RIDGE POLES Pcs 6 1,000 2.30 13.79
2 RAFTER POLES Pcs 30 800 1.84 55.17
3 WALL POST POLES Pcs 10 500 1.15 11.49
4 BATTENS Pcs 80 200 0.46 36.78
5 BLACK PLASTIC PAPER Part Roll 1 6,000 13.79 13.79
6 TIE WIRE Roll 1 2,000 4.60 4.60
7 3” NAILS Kg 2 1,000 2.30 4.60
8 TIMBER FOR DOOR (Inc fittings) Pcs 1 6,000 13.79 13.79
9 TIMBER FOR WINDOWS (Inc fittings) Pcs 2 2,000 4.60 9.20
10 EARTH BRICKS Pcs 2,400 3 0.01 16.55
11 THATCH Pcs 1 9,000 20.69 20.69

88 www.shelterprojects.org SHELTER PROJECTS 2015 - 2016