7.3.2. Ecosystem service maps 
in agriculture
Louise Willemen, Sarah Jones, 
Natalia Estrada Carmona & Fabrice DeClerck
Introduction
Agricultural ecosystems are the largest eco- ES in agricultural landscapes operate across 
systems in the anthropocene. To produce different spatial and temporal levels: before 
food, fodder and fuels, these agricultural an ES reaches the field, it may have moved 
systems strongly depend on a reliable flow of over various distances from different land 
ecosystem services; examples include water, cover types in the surrounding areas. For ex-
pollination, pest control, soil fertility and the ample, soil conservation practices on slopes 
gene pool of wild crop relatives. At the same reduce the negative impact of sedimentation 
time, it is well known that many agricultur- or landslide risk on the downslope. Under-
al practices and the expansion of agricultural standing this multi-level aspect (where ES 
areas are a major threat to well-functioning come from and flow to and at what point in 
healthy ecosystems. However, the inverse can time) is crucial for an effective management 
arguably be just as true; agriculture, if well of ES flows in rural areas. 
managed, can become an important means 
by which to secure and safeguard ecosystem In this chapter, we reflect on the role of spa-
services (ES). Agriculture has been the most tial information on ES for the sustainable 
direct way humans altered their natural sur- management of agricultural areas. The use 
roundings and has brought major increas- and selection of ES to consider and their 
es in well-being and income to humans. It mapping approaches depend on: i) the 
is important to realise that most ES result strength of the relationship between agricul-
in human benefits only after human input tural production systems and ES supply and 
or activities, such as seeding and harvesting ii) the spatial extent of the supply, flow and 
crops, travelling to attractive locations, or re- management level of the ES.
directing water (Chapter 5.1). 
Agricultural systems are intensely managed Ecosystem services and 
by humans and are more controlled and reg-
ulated than most other ‘ecosystems’. Many agricultural production links
governance systems are in place to manage 
and distribute excludable and rival goods In 2014, The Economics of Ecosystems and 
(e.g. water board for irrigation water, fishing Biodiversity initiative (TEEB) initiated a 
quota, timber extraction licences). This high specific study on the value of ES and bio-
level of human management and regulation diversity across agricultural systems: TEEB 
creates opportunities for securing and safe- for Agriculture and Food (TEEBAgFood). 
guarding ES for agriculture and non-agri- TEEBAgFood has identified the positive 
cultural production uses. (provisioning and regulating services) and 
Chapter 7 319
negative (environmental impacts) flows to ample, the supply of the ES ‘nutrient cycling’ 
and from agricultural systems. The quanti- is particularly relevant for low input farming 
fication of these services helps to assess the systems. In contrast, closely managing nutri-
dependence and impact of production sys- ent cycling via an ES based approach is not as 
tems on ES supply. relevant on farms where this is provided by 
Figure 1. Linkages between ES and agricultural management types for ES production, ES dependence 
and ES impact per spatial level. The white arrows indicate to which farming type the ES relate, from low 
to high input.
However, not all ES have equal relevance for synthetic fertilisers. In Figure 1, this is shown 
all farming systems. In Figure 1, we show by the arrow indicating the lower input farm-
the assumed and simplified link for high to ing systems only for this ES. Some ES are rel-
low input farming systems to relevant ES evant for all farming systems: all farms will 
based on their supply, ES dependence and produce food, fodder or fuel crops, they all 
ES impact. The figure also shows on which rely on specific water and climate conditions 
spatial level these interactions take place and and all conversions of land to agriculture will 
therefore need to be managed. “Input” refers impact the natural habitat.
here to pesticides, fertilisers and water (not  
to labour or machinery). The white arrows in Figure 1 could be used as a general guide for 
this figure indicate the farming systems for selecting the specific ES to be mapped, in ad-
which the specific ES (and thus information dition to the location-specific ES information 
on this ES) is relevant. The general assump- needs and focus. Maps of ES play an import-
tion is that low input farms are more depen- ant role in land management for: the assess-
dent and have less impact on ES compared ment of the current state of ES in rural ar-
to conventional high input farming. For ex- eas, impact analyses of agriculture on ES and 
320 Mapping Ecosystem Services
the monitoring of ES to support sustainable ment, farm and field level maps alone are 
management of agricultural areas. Land man- insufficient, as agriculture mostly supplies, 
agement, as well as the generation of spatial impacts and depends on ES from larger spa-
information, has so far mostly focused on the tial extents. The spatial extent of ES and the 
ES supply (agricultural goods) and ES impact related mapping requirements (data resolu-
(e.g. environmental impact assessments) and tion, accuracy) are described in Chapter 5.2.
less so on the enabling of common public 
goods on which ES depend (central blue bar 
of Figure 1). The TEEBAgFood project calls Applications of ES mapping in 
these the ‘invisible’ positive flows. Maps can 
make these invisible flows ‘visible’, facilitating agricultural areas
their inclusion in decision-making.
Current work demonstrates that ES maps 
and the process of generating maps can 
Ecosystem service maps for address important land management ques-
farms and beyond tions in agricultural areas across the globe. 
Studies have shown that the process of map-
ping ES as well as the maps themselves can 
Decisions on agricultural practices are typi- be used to: i) visualise the scales at which 
cally made at farm level. However, most ES different services operate; ii) assess locations 
on which agriculture depends and impacts of ES supply and beneficiaries highlighting 
often have a spatial level exceeding the farm. dependencies; iii) visualise impacts which 
Figure 1 shows that difference: few ES are are often considered invisible externalities of 
purely linked to field level, while many ES agriculture, both positive and negative; iv) 
are related to the ‘full eco-agri-system’ which facilitate negotiations amongst stakeholders, 
can cover landscapes, watersheds or even the including payment schemes and v) target 
global system depending on the ES in ques- intervention locations required to ensure or 
tion. Thus, when mapping ES to support improve ES supply. An example of this type 
decision-making in agricultural manage- of ES mapping study is presented in Box 1. 
Box 1 . Managing reservoir catchments to secure transboundary 
ES delivery in the Volta basin
The Volta River flows through six West African countries, draining a 407,000 km2 area that is home to 
over 20 million people. The Volta basin is subject to highly variable rainfall, yet timely supply of a sufficient 
quantity of quality water is essential for the rural households that rely on crop, fish or livestock production 
for their livelihood. Over 1000 small and several large dams have been constructed in the basin since the 
1950s to help maintain a year-round supply of agricultural water. Ecosystem processes in the reservoir 
catchments provide a service for reservoir-users by regulating the quality, quantity and timing of reservoir 
water supplies, making the network of land-users, reservoir systems and water beneficiaries tightly inter-
connected. Bioversity International and its partners are working with smallholder farmers and local and 
regional government in the Volta basin to facilitate evidence-based ES management decisions. Many of 
these stakeholders identify soil erosion and associated sedimentation as a key threat to reservoir water sup-
plies and water management authorities are seeking to minimise erosion through improved management 
of land adjacent to the stream network. The ES model WaterWorld , is used here to investigate the effect 
on water supply and the control of soil erosion rates by ensuring: 1) 100 % herbaceous plant cover and 2) 
Chapter 7 321
100 % tree cover, on land within 100 m of waterways in dam catchments across the Volta basin. Results 
indicate that targeting herbaceous vegetation cover in riparian zones (Scenario 1) would be more effective 
than targeting tree cover (Scenario 2) for improving water availability, although benefits are unevenly dis-
tributed across the region and generally higher in the south. Local variations in annual water balance are 
expected particularly under the tree cover scenario, with the annual water supply falling to less than half 
of its baseline level (a decrease of more than 100 %) in several dispersed locations across the region. The 
area, highlighted in the annual water supply inset maps below, illustrates that water supplies are generally 
expected to decrease on the Burkinabé side of the border under both scenarios while, on the Ghanaian 
side, water balance is expected to increase by up to 10 % or more in most places under herbaceous cover 
(Scenario 1), but continue to fall under tree cover (Scenario 2). The difference in water supply results 
between the scenarios can be largely explained by a difference in evapo-transpiration losses which will be 
higher from tree cover than herbaceous cover. In contrast, both vegetation types appear to be effective at 
controlling sediment. Both scenarios indicate erosion control rates adjacent to waterways will increase 
across the basin where there is perennial vegetation cover, with the largest erosion prevention impacts 
occurring near the headwaters of the stream network where slopes are steepest. The erosion control inset 
maps below illustrate that reduced erosion rates may be up to 100 % compared to baseline levels in some 
areas. The model outputs show that ensuring year-round vegetation cover on land adjacent to waterways, 
particularly with herbaceous plants and near stream headwaters, could be an effective strategy to control 
sedimentation rates and improve regional water supplies. Much of this riparian land is currently used 
for crop and livestock production and restricting agriculture on this land would negatively impact on 
thousands of smallholder farmers. Careful management of vegetation cover on existing agricultural land 
combined with protection and restoration of natural vegetation in adjacent areas could represent a viable 
option for implementing a riparian management scheme with minimal losses to food production. This 
would mean agricultural land in riparian zones is selectively managed to ensure year-round plant cover by, 
for example, using perennial species such as bananas, perennial rice and cover crops, while natural vegeta-
tion is restored and protected on adjacent non-agricultural land.
Mapping relative changes in ecosystem servces across the Volta basin under two riparian buffer 
management scenarios.
Scenario 1: Herbaceous plant cover (natural, crops, cover crops) in 100 m buffer along waterways in dam watersheds.
Change from baseline (%)
-1000% - -100%
Main map scale: 
-99% - -11% 1:17,000,000. 
-19% - -1% Minor map scale: 
1:5,000,000. 
0% (no change) Data sources: 
1% - 10% GAUL (admin bounderies); 
GRUMP (settlements) 
11% - 100% WaterWourld V2 - 
KCL/AmbioTEK 
101% - 1,000% (all other data)
322 Mapping Ecosystem Services
Scenario 2: Tree cover (natural, orchards, plantations) in 100 m buffer along waterways in dam watersheds.
Change from baseline (%)
-1000% - -100%
Main map scale: 
-99% - -11% 1:17,000,000. 
-19% - -1% Minor map scale: 
1:5,000,000. 
0% (no change) Data sources: 
1% - 10% GAUL (admin bounderies); 
GRUMP (settlements) 
11% - 100% WaterWourld V2 - 
KCL/AmbioTEK 
101% - 1,000% (all other data)
Further reading
Fremier AK, Declerck FAJ, Bosque-Pérez NA, Poppy GM, Chiotha S, Eigenbrod F, Harvey 
Carmona NE, Hill R, Joyal T, Keesecker CA, Honzák M, Hudson MD, Jarvis A, 
L, Klos PZ, Martínez-Salinas A, Niemey- Madise NJ, Schreckenberg K, Shackleton 
er R, Sanfiorenzo A, Welsh K, Wulfhorst CM, Villa F, Dawson TP (2014) Food secu-
JD (2013) Understanding Spatiotemporal rity in a perfect storm: using the ecosystem 
Lags in Ecosystem Services to Improve In- services framework to increase understand-
centives. BioScience 63: 472-482. ing. Philosophical Transactions of the Royal 
Society B: Biological Sciences: 369.
Mulligan M (2013) WaterWorld1: a self-pa-
rameterising, physically based model for Power AG (2010) Ecosystem services and 
application in data-poor but problem-rich agriculture: tradeoffs and synergies. Phil-
environments globally. Hydrology Re- osophical Transactions of the Royal Soci-
search 44(5): 748. ety of London B: Biological Sciences 365: 
2959-2971.
TEEB (2015) TEEB for Agriculture & Food: 
an interim report. United Nations Environ-
ment Programme, Geneva, Switzerland.
1  www..org/waterworld 
Chapter 7 323