ArticlePDF Available

Abstract and Figures

Southern Africa is on its second bird atlas project, with the second starting 16 years after fieldwork for the first was completed. The paradigm of the second atlas project has evolved so that it aims to monitor bird species in both space and time, producing a “movie” rather than a “snap-shot” of distribution. The second project uses a five-minute geographical grid, generating spatial units known as “pentads”: these are 9.2 km × 9.2 km on the equator, with the east-west component reducing to 8.3 km at the latitude of South Africa. The paper describes the fieldwork protocol. Fieldwork is undertaken throughout the year, so that studies of the time of migration are feasible. By December 2016, nine million records of bird distribution had been collected by citizen scientists, and there was data available for 77 % of the 17,339 pentads in South Africa, Lesotho and Swaziland. Projects using the same protocol have started in Kenya and Nigeria. Two sets of products comparing the databases of the first and second bird atlas projects are starting to emerge: studies of how ranges have changed between through time; studies of how the timing of migration has changed over a period as short as two decades. The especial value of this initiative to European ornithology is the information on the distribution and timing of arrival and departure of migrants to Africa.
Content may be subject to copyright.
64 VOGELWELT 137: 64–70 (2017)
1. Introduction
e First Southern African Bird Atlas Project (SABAP1)
(H et al. 1997a, b) provided a snap-shot of bird
distributions through the 1980s and up to 1991. Aer a
16-year gap, SABAP2 was initiated in July 2007, and is
ongoing (H et al. 2008, U & B
2016a).
Usually, the objective of a bird atlas project is to
provide a snap-shot of bird distributions, taken over a
period of years. is, certainly, was the explicit goal of
the First Southern African Bird Atlas Project (SABAP1)
(the “time exposure” was primarily the 1987–91 eld-
work period), and this was the initial goal of the Second
Southern African Bird Atlas Project, which started in
2007 and is ongoing (H et al. 2008). However,
SABAP2 has shied to a new paradigm, and visualizes
itself as producing the movie of bird distributions, with
annual increments (U 2016a). It is probably
the rst atlas project to cast itself into a long term moni-
toring role, with the objective of tracking bird distri-
butions both in time and in space. It is trying to track
changes in distribution in real time.
SABAP1 included the countries Botswana, Lesotho,
Namibia, South Africa, Swaziland and Zimbabwe; most
of the eldwork was undertaken during 1986–91, but
it included compatible data from 1980 onwards. It col-
lected a total of 7.3 million records of bird distribution.
e results of SABAP1 were published in two volumes
(H et al. 1997a, b). e spatial grid was the 15
minutes of latitude by 15 minutes of longitude, the so
called quarter degree grid cell. SABAP2 initially was
conned to South Africa, Lesotho and Swaziland. e
spatial grid was 5 minutes of latitude by 5 minutes of
longitude, generating grid cells which are known as
pentads, nine pentads to a quarter degree grid cell
(H et al. 2007).
is paper aims to describe the key features of the
SABAP2 project, under three broad headings: proto-
col – how the eldwork is done; process – how the
project works in practice; product – what the project
has achieved to date. It goes on to assess the potential
of expansion of the systems to Africa as a whole.
2. Protocol
e grid system that underpins the protocol generates
spatial units of ve geographical minutes of latitude
north-south and ve minutes of longitude east-west.
ese grid cells are known as “pentads” and there are
144 pentads per one degree cell. At the latitude of
South Africa, the dimensions of a pentad are 9.2 km ×
8.3 km, and on the equator they are 9.2 km × 9.2 km.
U (2016a) described the motivation for this
e Second Southern African Bird Atlas Project:
protocol, process, product
Les G. Underhill, Michael Brooks & Megan Loie-Eaton
Underhill L. G., M. Brooks & M. Loie-Eaton 2017: e Second Southern African Bird Atlas
Project: protocol, process, product. Vogelwelt 137: 64–70.
Southern Africa is on its second bird atlas project, with the second starting 16 years aer eldwork
for the rst was completed. e paradigm of the second atlas project has evolved so that it aims
to monitor bird species in both space and time, producing a “movie” rather than a “snap-shot”
of distribution. e second project uses a ve-minute geographical grid, generating spatial units
known as “pentads”: these are 9.2 km × 9.2 km on the equator, with the east-west component
reducing to 8.3 km at the latitude of South Africa. e paper describes the eldwork protocol.
Fieldwork is undertaken throughout the year, so that studies of the time of migration are feasi-
ble. By December 2016, nine million records of bird distribution had been collected by citizen
scientists, and there was data available for 77 % of the 17,339 pentads in South Africa, Lesotho
and Swaziland. Projects using the same protocol have started in Kenya and Nigeria. Two sets
of products comparing the databases of the rst and second bird atlas projects are starting to
emerge: studies of how ranges have changed between through time; studies of how the timing
of migration has changed over a period as short as two decades. e especial value of this initia-
tive to European ornithology is the information on the distribution and timing of arrival and
departure of migrants to Africa.
Keywords: bird atlas project, citizen science, Africa, project protocol
VOGELWELT 137: 64–70 (2017) 65
grid system, and the reason for its adoption. rough-
out Africa, biodiversity atlas projects are appropriately
based on a geographical grid, because the continent
does not stretch too far north or south for the tapering
of the east-west distance of grid cells to be problematic,
as occurs in Europe, where a UTM-based grid is needed
(H & B 1997).
For SABAP1 and SABAP2 eldwork is continuous
throughout the year. Unlike Europe, where the breeding
season is in general well-dened, breeding by birds in
most of Africa is fairly chaotic, driven by unpredictable
rain events, rather than by seasons. So the concept of
a “breeding atlas” is not a feasible concept. Even in the
Western Cape, in southwestern South Africa, which
has a fairly predictable Mediterranean climate with wet
winters and dry summers, there are species breeding
throughout the year (Het al. 1997a, b). ere
are passerine species which breed during the cold, wet
winter; the sunbirds and Cape Sugarbird Promerops
cafer are winter breeders. Many seabirds, including the
African Penguin Spheniscus demersus, breed in autumn,
winter and spring.
In broad brush terms, a “full-protocol” checklist for
SABAP2 consists of at least two hours of intensive eld-
work within the pentad, and an observer is not permit-
ted to start a new checklist for a pentad until ve days
have elapsed. Species are listed in the order in which
they are observed (U 2016a).
In more detail, a checklist is full-protocol if it con-
forms to the following standards (U 2016a).
(1) Atlasers operate within the boundaries of a pentad,
and record bird species observed within it. (2) ey
aim to make as complete a list as possible of the species
that were present in the pentad during the observation
period. ey do not need to cover the whole pentad,
but they need to try to sample as many of the habitats
that occur in the pentad as possible (acknowledging
that there are oen constraints on access). ey are
expected to consult the available maps and satellite
images, and get permission to access features (such as
wetlands) that are not accessible from a public road.
(3) Atlasers do focused birding for a minimum of two
hours in their pentads, but can continue for longer,
and are encouraged to do so if they are still regularly
adding species. (4) ey list the species in the order
in which they see the species, on the understanding
that the most abundant species in a pentad will tend
regularly to be among the rst species to be recorded.
(5) ey can continue to add “additional species” to
their list for ve days. (6) en they can start a new
list for the pentad. e ve-day gap helps ensure that
each successive list submitted by the same observer is
not simply a clone of the previous list, which would
be the case if observers were allowed to submit lists
for the same pentad on successive days. A checklist
produced in this way is referred to as a “full-protocol
checklist”.
Atlasers are encouraged also to submit “ad hoc check-
lists” and especially to do this in areas that are poorly
covered. ese checklists are most frequently made in
periods of less than an hour. Sometimes, they are made
over multiple days at a single spot, and therefore refer to
only to a limited section of the pentad, and are not rep-
resentative of all the habitats that occur in the pentad.
Various systems have been designed to capture data
into the SABAP2 database. By 2016, the dominant
data capture system was a mobile phone app known
as BirdLasser (http://birdlasser.com) (N et al. 2016).
is app eliminates the need for advanced map-reading
skills, and the list generated on the phone while atlasing
becomes the list submitted to the project, straight from
the mobile phone, eliminating the need for subsequent
data capture.
e database is continuously updated with the
incoming data. Distribution maps, for example, are
available from the live database on demand, and make
use of all data. By December 2016, nine million records
of bird distribution had been captured, and full protocol
checklists were available for 77 % of the 17,339 pentads
in Lesotho, South Africa and Swaziland.
3. Process
ere are two components to this section, one physical,
one psychological.
Physically, the SABAP2 data are stored in an enter-
prise level database, mysql. All data are backed up daily,
and archives are kept for a period of two years. Each
record on every submitted checklist is vetted for valid-
ity against the current database, as well as historical
datasets where available. Any record which is deemed
to be out of the known range or season of the species
are queried with the observer. is is done by soware,
and the queries, known as “out of range forms” (ORFs)
arrive within seconds of submission. All SABAP2 sys-
tems run on open source soware wherever possible,
with any additional requirements met by custom writ-
ten solutions.
e unexpected processes surrounding the SABAP2
protocol are related to psychology. e alternative to
having a grid system underlying the protocol, as with
the SABAP2 protocol, is to have the naïve protocol
(U 2016a). is protocol simply says “georef-
erence the positions of as many birds as possible (and let
the clever analysts work out what to do with the data)”.
With the SABAP2 protocol there is a precise sense of
entering the pentad, and starting eldwork. e atlaser
remains within the boundaries of the pentad and strate-
gizes to record as many dierent species within the pen-
tad as feasible. It is this strategizing which introduces
the concept of gamication (persuasive motivation)
into the eldwork. ere is a sense of working against
the clock which has the positive eect of sharpening
the senses. ere is a precise moment of leaving the
66 L. G. UNDERHILL et al.: e Second Southern African Bird Atlas Project
pentad and completing the eldwork. is generates
a sense of accomplishment which the naïve protocol
cannot provide. Gamication is further described by
A & U (2017). With the naïve protocol
there is no sense of a predetermined start and nish.
e SABAP2 protocol provides a precise mechanism
for evaluating how much eldwork has been done in
a region: measured in either number of checklists
submitted for the pentad or the region, or numbers
of hours spent doing intensive eldwork. e pentad
system spawns the crucial concept of the “coverage
map, clearly showing the areas for which no data exists,
shown as pentads with no data on the coverage map.
It is only the SABAP2 protocol which could generate
the vulgar slang which is used to describe the process
of doing the rst checklist for a pentad: “I did a virgin
pentad.” However unsatisfactory the metaphor, it is
profoundly motivating.
e coverage map is the ever-visible tool for planning
atlasing expeditions to poorly covered regions (A-
 & U 2017). e coverage map becomes
the catalyst for trip planning, even for family holidays.
An unvisited pentad surrounded by pentads with data
is perceived by atlasers as a blot on the landscape, and
it becomes a personal challenge to work out the access
strategy to reach it.
Psychologically, the grid system helps to standardize
eort. e division of the landscape into pentads pro-
vides the critical guidance about how far you need to
travel before the next Pied Crow Corvus albus becomes
important. Once you cross the boundary into a new
pentad (and the BirdLasser app takes care of this,
wherever you are in Africa) you start a new list. e
search for all species is reignited. e pentad grid sys-
tem relentlessly provides this common discipline on all
participants, and also on a single participant through
the course of a day’s eldwork.
e coverage map represents gami-
cation at its best. Gamication has
nothing to do with turning atlasing
into a “game”; it is better dened
as “persuasive design” (A &
U 2017). It engages people
and helps motivate them to achieve
the goals of the bird atlas project.
Gamication taps into the basic natu-
ral desires of people: socializing, learning, competition,
achievement, status and altruism. For the bird atlas
project, it leverages participation through the recogni-
tion of individual achievements in attaining common
goals. Gamication succeeds because it makes chores
feel like games (A & U 2017).
By good fortune, it turns out that the choice of pen-
tad as sampling unit was inspired. Psychologically, it
turns out to be just the right size. It is large enough to
be challenging, but small enough to be manageable in
2–4 hours of birding.
4. Product
e database produced by a bird atlas project is not
really a product, but it is merely a stepping stone
towards actual products. One important category of
product is the written paper, mostly scientic papers
in journals; there is a published bibliography of papers,
books and theses which have depended fundamentally
on data out of the two bird atlas projects in southern
Africa (U 2016b).
is paper highlights four key products that have
emerged out of SABAP2: a new series of distribu-
tion maps, range change maps, analyses of timing of
migration, and detailed studies of single species. e
protocol is designed to facilitate a wide range of sta-
tistical analyses, which makes data analysis relatively
straightforward, and enables relatively small changes
to be detected (U 2016a). is is in sharp
contrast to data collected with the naïve protocol, for
which data analysis is complex.
e pentad scale enables high resolution distribu-
tion maps to be created (Fig. 1) (U & Brooks
2016b). e guidelines for the interpretation of these
distribution maps are complex (Box 1); the interpreta-
BirdMAP species distirbution map at pentad scale –
Artverbreitungskarte auf Pentade
Spotted Flycatcher – Grauschnäpper
Fig. 1: e distribution map for the
Spotted Flycatcher, which needs to be
interpreted using the insights contained
in Box 1. e grid cells are pentads, as
used for SABAP2. – Verbreitungskarte des
Grauschnäppers, die Interpretation sollte
auf Basis der in Box 1 formulierten Hilfen
erfolgen.
VOGELWELT 137: 64–70 (2017) 67
tion will be simplied once pentads have a minimum
of four checklists each, a concept which was a theme
of A & U (this volume). e example
plot for this paper is the Spotted Flycatcher Muscicapa
striata, a non-breeding migrant from the Palearctic
which is distributed widely across northeastern South
Africa, with most records coming from within the
savanna biome (Fig. 1). ere are also many records
across the southwestern two-thirds of South Africa; this
is a pattern shared with many long-distance migrants,
with a scatter of points south of the non-breeding range
adopted by the bulk of individuals of the species.
e critical ability aorded by the second atlas pro-
ject in a region is ability to detect changes in ranges.
U & B (2016c) pointed out that com-
parisons between SABAP1 and SABAP2 had proven
to be statistically complex. Because SABAP1 used a
quarter degree grid, the SABAP2 data for the pentads
in quarter degree grid cell need to be pooled. e map-
ping system devised by Underhill & Brooks (2016c)
made use of transformations which provide estimates
of the relative change in abundance in the two decades
between SABAP1 and SABAP2 (Fig. 2). In grid cells
shaded blue, dark green and light green the relative
increases in abundance were large, moderate and small
respectively. Red, orange and yellow represent relative
decreases in abundance which are large, moderate and
small respectively. U & B (2016c) pro-
vided a full discussion of the advantages and limitations
of these range-change maps. is map for the Spotted
Flycatcher suggests that there had been substantial
increases in abundance for this species in the north-
eastern region of South Africa between SABAP1 and
SABAP2. e near-continuous region of blue grid cells
for Gauteng and adjacent Mpumalanga suggest that
there had been large increases in abundance in this area
(Fig. 2). e explanation for this increase is not clear,
but it is believed likely to be due to bush encroach-
ment (L-E 2014). In the southwest of South
Africa, there appears to be a mixture of red and blue
grid cells (Fig. 2). What this particular representation of
change fails to show is whether the increase or decrease
has been o a low base; it only shows an estimate of
the relative change in abundance. Over southwestern
South Africa, records of Spotted Flycatcher were rela-
tively rare for both SABAP1 and SABAP2 (Fig. 1), so
from the perspective of relative change in abundance
the species has either gone extinct or colonized from
scratch a particular grid cell, which is then shaded red
or blue respectively (Fig. 2). us the range-change
maps need to be understood in terms of the distribu-
tion map of Fig. 1.
An important, and unexpected, by-product of the
continuous eldwork adopted for both SABAP1 and
SABAP2 is the ability to monitor the arrival, depar-
ture and regional movements of migrants. ere are
several classes of migrants in South Africa: migrants
which breed in the Palearctic, and which spend the
non-breeding season in the south; intra-African
migrants, which breed in the atlas region, and which
spend the non-breeding season farther north in Africa,
and altitudinal migrants, especially between the higher
lying areas of the interior (where the altitude is mostly
above 1,200 m above sea level) and the coastal plain
of KwaZulu-Natal (H et al. 1997a, b). Pub-
lished papers have examined and described changes
in the timing of migration for Palearctic migrants and
for intra-African migrants (A et al. 2012, B-
 et al. 2015).
We conclude this section by noting studies using bird
atlas data in single species studies. Both species are in
threat categories. B et al. (2014) demonstrated the
Fig. 2: e range-change map for the Spot-
ted Flycatcher. e map displays changes
in estimated relative abundance between
SABAP1 and SABAP2. e grid cells
are quarter degree grid cells, as used by
SABAP1. e colour coding is described
in the text. – Die Karte visualisiert Ver-
breitungsänderungen des Grauschnäp-
pers. Änderungen der geschätzten relativen
Abundanz zwischen SABAP1 und SABAP2.
Die Gitterfelder sind Viertelquadranten,
wie für SABAP1 verwendet. Gitterfelder
in blau, dunkelgrün und hellgrün weisen
jeweils auf starke, moderate und schwache
Zunahmen der relativen Abundanz hin. Rot,
orange und gelb repräsentieren jeweils eine
starke, moderate und schwache Abnahme
der relativen Abundanz.
Relative change in abundance between SABAP 1 & 2 –
relative Abundanzänderungen zwischen SABAP 1 & 2
Spotted Flycatcher – Grauschnäpper
68 L. G. UNDERHILL et al.: e Second Southern African Bird Atlas Project
Box 1: Interpretation guidelines for new generation SABAP2 distribution maps. – Interpretationshilfen für die neue
Generation der SABAP2 Verbreitungskarten.
1. e cells on the maps are pentads, 5 minutes of latitude by 5 minutes of longitude, 9.2 km north-south × c.8.3 km
east-west. – Die Zellen auf den Karten sind Pentaden, 5 Längengradminuten zu 5 Breitengradminuten, 9,2 km Nord-Süd
x ca. 8,3 km Ost-West.
2. ere are two shading systems, one for pentads with less than four full protocol checklists, and another for pentads with
four or more full protocol checklists. e rst system shows presence-absence, the second shows reporting rate. – Es
gibt zwei Schattierungssysteme, eins für Pentaden mit weniger als vier vollständige Begehungen/Erfassungsdurchgängen
und ein weiteres für Pentaden mit vier oder mehr vollständigen Begehungen. Das erste System zeigt Präsenz/Absenz
Informationen, das zweite die Melderate.
3. If there are less than four checklists, there are three alternatives: turquoise = no data at all for the pentad; white circle
= species not recorded, although there is some data for the pentad, consisting of between one record (incidental or ad
hoc) and three full protocol checklists; grey = species demonstrated to be present in pentad. e white circles can be
interpreted as possibly absent. – Bei weniger als vier Begehungen gibt es drei Alternativen: türkis = keine Daten für die
Pentade; weißer Kreis = Art nicht festgestellt, obwohl Daten für die Pentade vorliegen, die aus einer (zufällig oder ad hoc)
oder bis zu drei vollständigen Begehungen beruhen; grau = Art in der Pentade nachgewiesen. Weiße Kreise können als
„wahrscheinlich abwesend“ interpretiert werden.
4. If there are four or more checklists, the reporting rates are represented in colour. – Bei vier oder mehr Begehungen wird
die Melderate der Arten über verschiedene Farben dargestellt.
If the species has not been recorded, the entire pentad is shaded white, and the species is probably absent. – Wenn eine
Art nicht festgestellt wurde, bleibt die Pentade weiß und die Art kommt dort wahrscheinlich nicht vor.
e reporting rates from the remaining pentads are sorted from smallest to largest, and split into six groups, which are as
even in size as possible. e “cut points” for the groups vary with the species. – Die Melderaten der verbliebenen Pentaden
sind von klein nach groß geordnet und in sechs möglichst ähnliche Gruppen unterteilt.
e pentads with the largest one-sixth of reporting rates are shaded dark blue, indicating the core of the range of the
species. e next sixths are shaded light blue, then dark green, the light green, then orange and nally yellow, for the
smallest one-sixth of reporting rates, where the species is most rarely recorded. – Die Pentaden mit dem höchsten Sechstel
der Melderaten (also den größten Anzahlen von Ind. der jeweiligen Art) sind dunkel blau schattiert und visualisieren das
Kernverbreitungsgebiet einer Art. Die weiteren Sechstel sind hellblau, dunkelgrün, hellgrün, dann orange und schließlich gelb,
für das Sechstel der geringsten Melderaten, wo eine Art nur selten angetroen wurde, dargestellt.
If a pentad has four or more checklists, and the species has only been recorded as an incidental or on an ad hoc list, then
the pentad is shaded yellow. – Wenn in einer Pentade mehr als vier standardisierte Begehungen erfolgten, eine Art aber nur
zufällig oder als Teil einer ad hoc Artenliste erfasst wurde, wird dies über ein gelbe Schattierung dargestellt.
e pentads shaded blue, either light or dark, represent the third of the range where reporting rates are largest, the pentads
shaded green, either light or dark, show the middle third of reporting rates, and yellow-orange pentads represents the third
of the range with the smallest reporting rates. Pentads shaded dark blue, light blue and dark green all have reporting rates
above the median reporting rate for the species. Half of the pentads are shaded these three colours. – Die entweder hell-
oder dunkelblau schattierten Pentaden repräsentieren das Drittel der Verbreitung mit den größten Melderaten, die hell- oder
dunkelgrün schattierten Pentaden stellen das mittlere Drittel der Melderaten dar und gelb-orangene Pentaden visualisieren
das Drittel der Verbreitung mit den geringsten Melderaten.
e reporting rate values for the ve cutpoints are only of academic interest. ey are deliberately not presented, because
they are not comparable between species. – Die Werte der Melderaten die hinter den fünf Trenngrenzen zwischen den
Häugkeitskategorien liegen sind nur von akademischem Interesse und werden absichtlich nicht genannt, da diese für unter-
schiedliche Arten nicht vergleichbar sind.
importance of the large protected areas for the Southern
Ground Hornbill Bucorvus leadbeateri, and H
et al. (2014) demonstrated that the decrease in abun-
dance of Secretarybirds Sagittarius serpentarius had
been precipitous and attributed this to bush encroach-
ment (L-E 2014). In both studies, the bird
atlas data were pivotal in highlighting and explaining
the plight of the study species. Conservation action is
being initiated.
5. Pan-African perspective
With an area of 30 million km2, Africa has about 400,000
pentads (U 2016a). Apart from the four larg-
est countries (Algeria, Democratic Republic of Congo,
Sudan and Libya), the remaining 50 countries on the
continent have fewer than 18,000 pentads, and most
have less than 10,000. On a country-by-country basis,
Africa is manageable on a pentad scale. e bird atlas
VOGELWELT 137: 64–70 (2017) 69
7. References
A, J. & L. G. U 2017: Gamication (persua-
sive design) in the Southern African Bird Atlas Project
(SABAP2). Vogelwelt 137: 19–22.
A, R., K. B, B. E, P. B, G.F. M
& L. G. U 2012: Novel methods reveal shis in
migration phenology of Barn Swallows in South Africa.
Proc. Roy. Soc. Lond. B 279: 1485–1490.
B, K. M., D. S. J, R. A & L. L. C
2014: Spatial occupancy models applied to atlas data
show Southern Ground Hornbills strongly depend on
protected areas. Ecol. Appl. 24: 363–374.
B, E. M. S., L. G. U & R. A 2015:
Patterns of bird migration phenology in South Africa
suggest northern hemisphere climate as the most consist-
ent driver of change. Glob. Chang. Biol. 21: 2179–2190.
project database at the Animal Demography Unit already
holds records for 23,000 pentads in Africa; that is 5.5 %
of the total. is kind of thinking is already gamication,
“persuasive motivation” (A & U 2017).
U & B (2016d) set out a vision
for atlasing in those parts of Africa where there is a
poor observer network. ey envisaged four sources
of data: (1) “full-protocol” checklists, with at least two
hours of eldwork, and a concerted eort to nd as
many species as feasible within a pentad; (2) “ad hoc”
checklists, usually made over periods shorter than an
hour, which provide a sample of the species of the pen-
tad; (3) photographic records, uploaded to the BirdPix
section of the ADU Virtual Museum (http://vmus.adu.
org.za) (U et al. 2016); and (4) historical
lists, in which the area covered is well dened, and
suciently small that it can be allocated to a pentad.
e inclusion of photographic records as a data source
enables people who cannot condently identify spe-
cies to participate. Each of the sections of the ADU
Virtual Museum has an “expert panel” which does
the formal identications. In some photographs, the
species is not identiable, but for most species even
a poor photograph is good enough (U &
N 2016).
e countries adjacent to South Africa have atlas
projects in operation. Farther north, Kenya Bird Map
is at 12 % coverage, and is growing rapidly. e Nigerian
Bird Atlas commenced in 2016, and is already at 3 %
coverage (T et al. 2016).
From a European perspective, expansion of this bird
atlas network to further countries in Africa is a crucial
component of developing the conservation priorities
for many species. A knowledge of non-breeding areas
is critical to understanding the life-cycles of the large
number of bird species breeding in Europe and migrat-
ing to Africa to evade the northern winter. e year-
round protocol used in Africa will facilitate studies of
timing of migration of these species, and how this is
changing in response to climate change.
6. Zusammenfassung
Underhill L. G., M. Brooks & M. Loie-Eaton 2017: Das zweite südafrikanische Vogelatlasprojekt: Protokoll, Prozess,
Produkte. Vogelwelt 137: 64–70.
In Südafrika laufen die Arbeiten am zweiten Vogelatlas des
Landes. Die Arbeiten begannen 16 Jahre nach Beendigung
der Feldarbeiten für den ersten Vogelatlas Südafrikas. Die
Paradigmen im Hinblick auf die Atlaserstellung haben sich
zwischenzeitlich entwickelt und das zweite Atlasprojekt zielt
auf eine Monitoring der Vogelarten in Raum und Zeit, umso
eher einen „Film“ als eine „Momentaufnahme“ der Artver-
breitung zu generieren. Das zweite Atlasprojekt nutzt ein
geographisches 5-Minuten-Raster, das räumliche Einheiten
generiert, die als Pentaden bezeichnet werden. Diese Pen-
taden haben auf Höhe des Äquators eine Ausdehnung von
9,2 km x 9,2 km, die ost-west Ausdehnung reduziert sich auf
der geographischen Breite Südafrikas auf 8,3 km. Der Beitrag
beschreibt die Erfassungsmethoden zur Erstellung des Altas.
Die Feldarbeiten erstrecken sich über das ganze Jahr und
ermöglichen Untersuchungen im Hinblick auf den Vogelzug
und Zugzeiten. Bis Dezember 2016 wurden von Bürgerwissen-
schalern etwa neun Millionen Datensätze zur Vogelverbrei-
tung gesammelt, so dass bereits für 77 % der 17.229 Pentaden
in Südafrika, Lesotho und Swasiland Daten verfügbar waren.
Projekte mit gleicher Erfassungsmethodik wurden zudem in
Kenia und Nigeria gestartet. Zwei Arten von Ergebnissen,
die die Datenbanken des ersten und zweiten Atlasprojektes
vergleichen bilden sich momentan heraus: Studien zu Ver-
breitungsänderungen zwischen den Erfassungsperioden und
Studien zu zeitlichen Veränderungen des Vogelzugs inner-
halb von nur zwei Dekaden. Der besondere Wert dieser Ini-
tiative für die europäische Ornithologie sind Informationen
zur Verbreitung sowie Ankuns- und Abzugszeiten für in
Afrika überwinternde europäische Brutvögel. Zur Deutung
der Verbreitungskarte (Abb. 1) sei der Leser auf die ebenfalls
übersetzen Interpretationsregeln in Box 1 verwiesen.
H, E. J. M. & M. J. B 1997: e EBCC Atlas of
European Breeding Birds: eir Distribution and Abun-
dance. T. & A. D. Poyser, London.
H, D. M., L. G. U & M. B 2007:
Southern African Bird Atlas Project 2 – Instruction
Manual. Animal Demography Unit, University of Cape
Town, Cape Town.
H, J. A., D. G. A, L. G. U, M. H-
, A. J. T, V. P & C. J. B 1997a: e
Atlas of Southern African Birds. Vol. 1: Non-passerines.
BirdLife South Africa, Johannesburg.
H, J. A., D. G. A, L. G. U, M. H-
, A. J. T, V. P & C. J. B 1997b:
e Atlas of Southern African Birds. Vol. 2: Passerines.
BirdLife South Africa, Johannesburg.
70 L. G. UNDERHILL et al.: e Second Southern African Bird Atlas Project
H, J. A., L. G. U & P. B 2008: e
seminal legacy of the Southern African Bird Atlas Project.
S. Afr. J. Sci. 102: 82–84.
H, S. D., C. T. S & L. G. U 2014: Sec-
retarybird Sagittarius serpentarius population trends and
ecology: insights from South African citizen science data.
PLoS ONE 9: e96772. doi: 10.1371/journal.pone.0096772.
L-E, M. 2014: Geographic Range Dynamics of
South Africa’s Bird Species. MSc thesis, University of
Cape Town, South Africa. Available at: http://www.adu.
uct.ac.za/sites/default/les/image_tool/images/352/stu-
dents/Loie-Eaton_M_2014_MSc_thesis.pdf
N, H., J. A & L. G. U 2016: e BirdLasser
app as a tool for gathering bird atlas data. Biodivers. Obs.
in press.
T, T., S. I & U. O 2016: Nigeria Bird Atlas
Project: How far so far? Progress report August 2016.
Biodivers. Obs. 7.50: 1–3. Available at: http://bo.adu.org.
za/content.php?id=243
U, L. G. 2016a: e fundamentals of the SABAP2
protocol. Biodivers. Obs. 7.42: 1–12. Available at: http://
bo.adu.org.za/content.php?id=235
U, L. G. 2016b: Bibliography: Research papers and
postgraduate theses which have been largely dependent
on data from the Southern African Bird Atlas Projects.
Biodivers. Obs. 7.43: 1–13. Available at: http://bo.adu.org.
za/content.php?id=235
U, L. G. & M. B 2016a: SABAP2 aer nine
years, mid 2007–mid 2016: coverage progress and priori-
ties for the Second Southern African Bird Atlas Project.
Biodivers. Obs. 7.37: 1–17. Available at: http://bo.adu.org.
za/content.php?id=235
U, L. G. & M. B 2016b: Pentad-scale distribu-
tion maps for bird atlas data. Biodivers. Obs. 7.52: 1–8.
Available at: http://bo.adu.org.za/content.php?id=245
U, L. G. & M. B 2016c: Displaying changes
in bird distributions between SABAP1 and SABAP2. Bio-
divers. Obs. 7.62: 1–13. Available at: http://bo.adu.org.za/
content.php?id=235
U, L. G. & M. B 2016d: e BirdMAP proto-
col: bird atlasing in Africa in regions where the observer
network is thin. Biodivers. Obs. 7.69: 1–5. Available at:
http://bo.adu.org.za/content.php?id=235
U, L. G. & R. A. N 2016: Fieldwork protocol
for the ADU Virtual Museum: How to photo hunt, aka
PHunt. Biodivers. Obs. in press.
Les G. Underhill, Michael Brooks, Megan Loie-Eaton, Animal Demography Unit, Department of Biological
Sciences, University of Cape Town, Rondebosch, 7701 South Africa; E-Mail: les.underhill@uct.ac.za
... SABAP2 maps bird species distributions using sightings data from several thousand volunteers. All bird species seen or heard are recorded following a defined protocol (Underhill et al. 2017) within a five minute of latitude × five minute of longitude grid cell (a 'pentad'). So-called 'full protocol' lists require observers to record species in the order encountered during sampling of all habitats in a pentad and spending at least two hours of active birding over a maximum of five days. ...
Article
The COVID-19 lockdown has had a marked effect on citizen science projects, such as the Southern African Bird Atlas Project, SABAP2. With their mobility severely constrained, most citizen scientists in South Africa were not able to submit full protocol lists during the Alert Level 5 lockdown. There was a 70% decline in full protocol list submissions in April 2020, and only 20% of the number of pentads was surveyed, compared with the same period in the previous three years. The decline in ad hoc lists is also evident, with about a 50% decline in the number of lists submitted and 25% of the usual number of pentads. The number of citizen scientists submitting ad hoc lists only declined by about 15% in April, compared with previous years, indicating that volunteers were still eager to contribute to the atlas project, despite the challenging conditions. The BirdLasser ‘Lockdown Challenge’ further encouraged participation. Although the ability of citizen scientists to contribute full protocol lists to SABAP2 may have suffered, other projects with a scope more suited to lockdown thrived. The ‘Lockdown Garden Surveys’ instigated as part of a study of urban bird communities proved extremely successful, with 283 people regularly contributing point counts, many on a daily basis. The overwhelmingly positive response to initiatives like these indicates that the lockdown may have actually increased people’s desire to participate in citizen science. The pandemic also has put the spotlight on nature within cities, encouraging people to take a fresh look at their surrounding greenspaces. The accompanying boom in citizen science and interest in nature is something that must be harnessed going forward, and we hope this can be sustained despite the harsh economic climate exacerbated by the pandemic.
... Birds in the UK have been particularly well monitored for decades and so our example was able to use high quality data for the whole of the area at a relatively fine spatial resolution. However, monitoring schemes for other taxa (Schmeller et al. 2009), and for birds in previously less well monitored areas (Underhill et al. 2017;Wotton et al. 2018), are becoming established in many parts of the world and will provide future opportunities for similar prioritization exercises. Advances in estimating abundance (Dennis et al. 2013) and new methods of monitoring at large spatial extents (Biggs et al. 2015) further increase the availability of suitable data for a range of taxa. ...
Article
Full-text available
Context Organisations acting to conserve and protect species across large spatial scales prioritise to optimise use of resources. Spatial conservation prioritization tools typically focus on identifying areas containing species groups of interest, with few tools used to identify the best areas for single-species conservation, in particular, to conserve currently widespread but declining species. Objective A single-species prioritization framework, based on temporal and spatial patterns of occupancy and abundance, was developed to spatially prioritize conservation action for widespread species by identifying smaller areas to work within to achieve predefined conservation objectives. Methods We demonstrate our approach for 29 widespread bird species in the UK, using breeding bird atlas data from two periods to define distribution, relative abundance and change in relative abundance. We selected occupied 10-km squares with abundance trends that matched species conservation objectives relating to maintaining or increasing population size or range, and then identified spatial clusters of squares for each objective using a Getis-Ord-Gi* or near neighbour analysis. Results For each species, the framework identified clusters of 20-km squares that enabled us to identify small areas in which species recovery action could be prioritized. Conclusions Our approach identified a proportion of species’ ranges to prioritize for species recovery. This approach is a relatively quick process that can be used to inform single-species conservation for any taxa if sufficiently fine-scale occupancy and abundance information is available for two or more time periods. This is a relatively simple first step for planning single-species focussed conservation to help optimise resource use.
Article
Large‐scale citizen‐science projects, such as atlases of species distribution, are an important source of data for macroecological research, for understanding the effects of climate change and other drivers on biodiversity, and for more applied conservation tasks, such as early‐warning systems for biodiversity loss. However, citizen‐science data are challenging to analyse because the observation process has to be taken into account. Typically, the observation process leads to heterogeneous and non‐random sampling, false absences, false detections, and spatial correlations in the data. Increasingly, occupancy models are being used to analyse atlas data. We advocate a dual approach to strengthen inference from citizen science data for the questions the programme is intended to address: (a) the survey design should be chosen with a particular set of questions and associated analysis strategy in mind and (b) the statistical methods should be tailored not only to those questions but also to the specific characteristics of the data. We review the consequences of particular survey design choices that typically need to be made in atlas‐style citizen‐science projects. These include spatial resolution of the sampling units, allocation of effort in space, and collection of information about the observation process. On the analysis side, we review extensions of the basic occupancy models that are frequently necessary with atlas data, including methods for dealing with heterogeneity, non‐independent detections, false detections, and violation of the closure assumption. New technologies, such as cell‐phone apps and fixed remote detection devices, are revolutionizing citizen‐science projects. There is an opportunity to maximize the usefulness of the resulting datasets if the protocols are rooted in robust statistical designs and data analysis issues are being considered. Our review provides guidelines for designing new projects and an overview of the current methods that can be used to analyse data from such projects.
Article
Full-text available
When the protocol for the Second Southern African Bird Atlas Project was developed in 2007, the committee involved had no inkling that they were designing an extremely effective motivational system. The term “gamification” had not yet been invented, and it is only recently that the gamification component of the protocol has been understood for what it really is. Gamification is an easily misunderstood term, and it means a design that in itself embodies motivation to achieve a set of goals. An alternative, and more intuitive term, is “stickiness.” A key element of gamification for the bird atlas project has proved to be the various forms of the “coverage map”. The paper provides examples of specific instances of gamification within the project, and it concludes by noting that the “stickiness” of the protocol has been a major factor in the success of the project.
Article
Full-text available
When the protocol for the Second Southern African Bird Atlas Project was developed in 2007, the committee involved had no inkling that they were designing an extremely effective motivational system. The term “gamification” had not yet been invented, and it is only recently that the gamification component of the protocol has been understood for what it really is. Gamification is an easily misunderstood term, and it means a design that in itself embodies motivation to achieve a set of goals. An alternative, and more intuitive term, is “stickiness.” A key element of gamification for the bird atlas project has proved to be the various forms of the “coverage map”. The paper provides examples of specific instances of gamification within the project, and it concludes by noting that the “stickiness” of the protocol has been a major factor in the success of the project.
Article
Full-text available
This paper recommends strategies for undertaking bird atlasing in Africa, making use of aspects of the SABAP2 protocol, in such a way that it is compatible with it. We call this protocol the BirdMAP protocol. The key element of compatibility is the five-minute grid system which generates cells known as pentads
Article
Full-text available
Data from two long-term citizen science projects were used to examine the status and ecology of a Red List species, the Secretarybird Sagittarius serpentarius (Vulnerable), in South Africa. The first phase of the Southern African Bird Atlas Project operated from 1987 until 1992, and the second phase began in 2007. The Coordinated Avifaunal Roadcounts (CAR) project began in 1993 and by 1998 had expanded to cover much of the south-eastern half of the country. Data submitted up until April 2013 were used. A new method of comparing reporting rates between atlas projects was developed. Changing reporting rates are likely to reflect changes in abundance; in this instance the data suggest that the Secretarybird population decreased across much of South Africa between the two atlas projects, with a widespread important decrease in the Kruger National Park. Habitat data from the CAR project were analysed to gain insight into the ecology of the species. Secretarybirds tended to avoid transformed habitats across much of the area covered by the CAR project. In the winter rainfall region of the Western Cape, which is characterised by heavily transformed fynbos vegetation, at least 50% of Secretarybirds recorded were in transformed environments. This implies that in the Fynbos biome, at least, Secretarybirds have adapted to transformed environments to some degree. However, in the rest of the country it is likely that habitat loss, largely through widespread bush encroachment but also through agriculture, afforestation, and urbanisation, is a major threat to the species. The methods developed here represent a new approach to analysing data from long-term citizen science projects, which can provide important insights into a species' conservation status and ecology.
Article
Full-text available
The first Southern African Bird Atlas Project was launched in 1986 and gathered bird distribution data from six countries of southern Africa. The project culminated with the publication of The Atlas of SouthernAfrican Birds in 1997. The database generated by the project, seven million bird distribution records, has been widely used by four groups: environmental consultants (for example, to locate electricity transmission lines), conservationists (planning conservation strategies), research scientists (especially macro-ecologists and biogeographers) and birders (ecotourism materials). By 2007, the database had spawned 50 research publications and eight Ph.D.s and master’s degrees. These products are a tribute to the more than 5000 ‘citizen scientists’, who gathered the bulk of the data. The atlas concept has been extended to frogs, reptiles, spiders and butterflies; a second bird atlas started in 2007 and will, for example, facilitate knowledge of the impact of environmental change on birds. The South African National Biodiversity Institute is playing a lead role in initiating these new projects.
Article
Full-text available
Many migratory bird species, including the barn swallow (Hirundo rustica), have advanced their arrival date at Northern Hemisphere breeding grounds, showing a clear biotic response to recent climate change. Earlier arrival helps maintain their synchrony with earlier springs, but little is known about the associated changes in phenology at their non-breeding grounds. Here, we examine the phenology of barn swallows in South Africa, where a large proportion of the northern European breeding population spends its non-breeding season. Using novel analytical methods based on bird atlas data, we show that swallows first arrive in the northern parts of the country and gradually appear further south. On their north-bound journey, they leave South Africa rapidly, resulting in mean stopover durations of 140 days in the south and 180 days in the north. We found that swallows are now leaving northern parts of South Africa 8 days earlier than they did 20 years ago, and so shortened their stay in areas where they previously stayed the longest. By contrast, they did not shorten their stopover in other parts of South Africa, leading to a more synchronized departure across the country. Departure was related to environmental variability, measured through the Southern Oscillation Index. Our results suggest that these birds gain their extended breeding season in Europe partly by leaving South Africa earlier, and thus add to scarce evidence for phenology shifts in the Southern Hemisphere.
Article
Current knowledge of phenological shifts in bird migration is largely based on data collected on Palearctic migrants at their breeding grounds; little is known about the phenology of these birds at their non-breeding grounds, and even less about that of intra-African migrants. Because climate change patterns are not uniform across the globe, we can expect regional disparities in bird phenological responses. It is also likely that they vary across species, since species show differences in the strength of affinities they have with particular habitats and environments. Here, we examine the arrival and departure of nine Palearctic and seven intra-African migratory species in the central Highveld of South Africa, where the former spend their non-breeding season and the latter their breeding season. Using novel analytical methods based on bird atlas data, we show phenological shifts in migration of five species - red-backed shrike, spotted flycatcher, common sandpiper, white-winged tern (Palearctic migrants) and diederick cuckoo (intra-African migrant) - between two atlas periods: 1987-1991 and 2007-2012. During this time period, Palearctic migrants advanced their departure from their South African non-breeding grounds. This trend was mainly driven by waterbirds. No consistent changes were observed for intra-African migrants. Our results suggest that the most consistent drivers of migration phenological shifts act in the northern hemisphere, probably at the breeding grounds. This article is protected by copyright. All rights reserved. This article is protected by copyright. All rights reserved.
Article
Determining the range of a species and exploring species--habitat associations are central questions in ecology and can be answered by analyzing presence--absence data. Often, both the sampling of sites and the desired area of inference involve neighboring sites; thus, positive spatial autocorrelation between these sites is expected. Using survey data for the Southern Ground Hornbill (Bucorvus leadbeateri) from the Southern African Bird Atlas Project, we compared advantages and disadvantages of three increasingly complex models for species occupancy: an occupancy model that accounted for nondetection but assumed all sites were independent, and two spatial occupancy models that accounted for both nondetection and spatial autocorrelation. We modeled the spatial autocorrelation with an intrinsic conditional autoregressive (ICAR) model and with a restricted spatial regression (RSR) model. Both spatial models can readily be applied to any other gridded, presence--absence data set using a newly introduced R package. The RSR model provided the best inference and was able to capture small-scale variation that the other models did not. It showed that ground hornbills are strongly dependent on protected areas in the north of their South African range, but less so further south. The ICAR models did not capture any spatial autocorrelation in the data, and they took an order, of magnitude longer than the RSR models to run. Thus, the RSR occupancy model appears to be an attractive choice for modeling occurrences at large spatial domains, while accounting for imperfect detection and spatial autocorrelation.
The EBCC Atlas of European Breeding Birds: Their Distribution and Abundance
  • E J M M J Hagemeijer
  • Blair
Hagemeijer, E. J. M. & M. J. Blair 1997: The EBCC Atlas of European Breeding Birds: Their Distribution and Abundance. T. & A. D. Poyser, London.
Geographic Range Dynamics of South Africa's Bird Species Available at: http://www.adu. uct.ac.za/sites
  • M H Loftie-Eaton
  • J Ainsley
Loftie-Eaton, M. 2014: Geographic Range Dynamics of South Africa's Bird Species. MSc thesis, University of Cape Town, South Africa. Available at: http://www.adu. uct.ac.za/sites/default/files/image_tool/images/352/stu- dents/Loftie-Eaton_M_2014_MSc_thesis.pdf Nel, H., J. Ainsley & L. G. Underhill 2016: The BirdLasser app as a tool for gathering bird atlas data. Biodivers. Obs. in press.
The fundamentals of the SABAP2 protocol
  • L G Underhill
Underhill, L. G. 2016a: The fundamentals of the SABAP2 protocol. Biodivers. Obs. 7.42: 1-12. Available at: http:// bo.adu.org.za/content.php?id=235