You are on page 1of 9

Contributed Paper

Use of Modified Threat Reduction Assessments to

Estimate Success of Conservation Measures within
and Adjacent to Kruger National Park, South Africa
Environmental Sciences & Policy Department, Central European University, Nádor u. 9, Budapest 1051, Hungary,

Abstract: The importance of biodiversity as natural capital for economic development and sustaining human
welfare is well documented. Nevertheless, resource degradation rates and persistent deterioration of human
welfare in developing countries is increasingly worrisome. Developing effective monitoring and evaluation
schemes and measuring biodiversity loss continue to pose unique challenges, particularly when there is a
paucity of historical data. Threat reduction assessment (TRA) has been proposed as a method to measure
conservation success and as a proxy measurement of conservation impact, monitoring threats to resources
rather than changes to biological parameters themselves. This tool is considered a quick, practical alternative
to more cost- and time-intensive approaches, but has inherent weaknesses. I conducted TRAs to evaluate the
effectiveness of Kruger National Park (KNP) and Limpopo Province, South Africa, in mitigating threats to
biodiversity from 1994 to 2004 in 4 geographical areas. I calculated TRA index values in these TRAs by using
the original scoring developed by Margoluis and Salafsky (2001) and a modified scoring system that assigned
negative mitigation values to incorporate new or worsening threats. Threats were standardized to allow
comparisons across the sites. Modified TRA index values were significantly lower than values derived from the
original scoring exercise. Five of the 11 standardized threats were present in all 4 assessment areas, 2 were
restricted to KNP, 2 to Limpopo Province, and 2 only to Malamulele municipality. These results indicate, first,
the need to integrate negative mitigation values into TRA scoring. By including negative values, investigators
will be afforded a more accurate picture of biodiversity threats and of temporal and spatial trends across sites.
Where the original TRA scoring was used to measure conservation success, reevaluation of these cases with the
modified scoring is recommended. Second, practitioners must carefully consider the need and consequences
of generalizing threats into generic categories for comparative assessments. Finally, continued refinement of
the methodology and its extension to facilitate the transfer of successful conservation strategies is needed.

Keywords: biodiversity threats, conservation success, Kruger National Park, Limpopo Province, threat reduction

Uso de Evaluaciones de Reducción de Amenaza Modificadas para Estimar el Éxito de Medidas de Conservación
Dentro de y Alrededor del Parque Nacional Kruger, África del Sur
Resumen: La importancia de la biodiversidad como capital natural para el desarrollo económico y sostén
del bienestar humano está bien documentada. Sin embargo, las tasas de degradación de los recursos y el
persistente deterioro del bienestar humano en los paı́ses en desarrollo es cada vez más preocupante. El
desarrollo de esquemas efectivos de monitoreo y evaluación y la cuantificación de la pérdida de biodiversidad
continúan presentando retos únicos, particularmente cuando existe escasez de datos históricos. La evaluación
de la reducción de riesgos (ERR) ha sido propuesta como un método para medir el éxito de la conservación
y como una medida del impacto de la conservación, que monitorea las amenazas a los recursos en lugar de
los cambios en los parámetros biológicos mismos. Esta herramienta es considerada una alternativa rápida
y práctica a métodos más costosos y tardados, pero tiene debilidades inherentes. Realice ERR para evaluar

Paper submitted September 24, 2007; revised manuscript accepted April 28, 2008.
Conservation Biology, Volume 22, No. 6, 1497–1505

C 2008 Society for Conservation Biology
DOI: 10.1111/j.1523-1739.2008.01030.x
1498 Threat Reduction Assessment

la efectividad del Parque Nacional Kruger (PNK) y la Provincia Limpopo, África del Sur, en la mitigación de
las amenazas a la biodiversidad de 1994 a 2004 en cuatro áreas geográficas. Calculé los valores del ı́ndice
de ERR en estas ERR mediante la clasificación desarrollada por Margoluis & Salafsky (2001) y un sistema
de clasificación modificado que asignó valores de mitigación negativos para incorporar nuevas amenazas o
empeorándolas. Las amenazas fueron estandarizadas para permitir comparaciones entre todos los sitios. Los
valores del ı́ndice de ERR modificados fueron significativamente menores que los valores derivados del ejerci-
cio de clasificación original. Cinco de las 11 amenazas estuvieron presentes en todas las áreas de evaluación,
dos estuvieron restringidas a PNK, dos a la Provincia Limpopo y solo dos a la municipalidad de Malamulele.
Estos resultados indican, primero, la necesidad de integrar valores de mitigación negativos a las ERR. Con
la inclusión de los valores negativos, los investigadores tendrán una imagen más precisa de las amenazas
a la biodiversidad y de las tendencias temporales y espaciales de los sitios. Se recomienda que los sitios
medidos con la clasificación de ERR original sean reevaluados con la clasificación modificada. Segundo, los
practicantes deben considerar cuidadosamente la necesidad y las consecuencias de generalizar las amenazas
en categorı́as genéricas para evaluaciones comparativas. Finalmente, se requiere el refinamiento continuo
de la metodologı́a y su extensión para facilitar la transferencia de estrategias de conservación exitosas.

Palabras Clave: amenazas a la biodiversidad, evaluación de reducción de amenazas, éxito de conservación,

Parque Nacional Kruger Provincia Limpopo

Introduction Threat reduction assessment (TRA) has been proposed

as an alternative evaluation tool to assess the success
The importance of biological diversity as natural resource of conservation interventions, especially where baseline
capital for economic development and sustaining human studies on biodiversity threats are absent (Salafsky &
welfare has been well documented (Ehrlich & Ehrlich Margoluis 1999; Margoluis & Salafsky 2001). Threat re-
1992; Costanza et al. 1997; Reaka-Kudla et al. 1997). Nev- duction assessments have been used to assess projects
ertheless, the rate at which natural resources continue to in Papua New Guinea, Indonesia, Madagascar (Salafsky
be degraded and the persistent deterioration of human & Margoluis 1999), Mexico (Hastings & Fischer 2001),
welfare in developing countries have caused concern at Kenya, Tanzania, and Uganda (Persha 2001; Mugisha &
local, national, and international levels (MEA 2005). This, Jacobson 2004), and are considered a quick, low-cost,
in part, has led to greater scrutiny and development of practical alternative to more cost- and time-intensive ap-
effective monitoring and evaluation systems (Noss 1990; proaches. Rather than monitoring changes to biological
Margoluis & Salafsky 1998; Teder et al. 2007). parameters themselves, TRA monitors threats to the re-
Nevertheless, monitoring the state of biodiversity is sources as a proxy measurement of conservation impact.
often costly, arduous, and time-consuming and has led It is sensitive to changes over short periods of time and
to the development of a range of conservation perfor- throughout a project site, allowing gross comparisons of
mance assessments by various institutions that focus on performance among projects at different sites (Margoluis
threats to biodiversity. These include the World Com- & Salafsky 2001). The key principle of TRA as an evalu-
mission on Protected Areas (WCPA) Framework for As- ation tool is that if threats to an area are mitigated, then
sessing the Management of Protected Areas (Hockings et the management will have succeeded. Conversely, if the
al. 2000), The Nature Conservancy’s (TNC) 5-S Frame- threats are not mitigated, the management approach will
work (TNC 2000, 2003), Wildlife Conservation Society’s have failed.
(WCS) Living Landscapes Program (WCS 2002), World Nevertheless, the TRA approach is not immune to bias.
Wide Fund for Nature’s (WWF) Rapid Assessment and Its category “percent threat reduced” is probably the
Prioritization of Protected Area Management (RAPPAM) biggest pitfall in this respect because, in addition to po-
(Ervin 2003a, 2003b), and conservation audits with the tential subjectivity on behalf of the participants (Salafsky
Conservation Measures Partnership’s (CMP) Open Stan- & Margoluis 1999; Mugisha & Jacobson 2004), TRAs in-
dards (CMP 2004). Each of these conservation perfor- corporate no scoring mechanism to allow for threats that
mance assessments has strengths and weaknesses on the had either arisen or worsened during the assessment pe-
basis of their feasibility, accuracy, and ease of utility (e.g., riod. In the TRA scoring (see Salafsky & Margoluis 1999),
Stem et al. 2005; Tucker 2005). Despite these tools, how- if a threat had not been addressed at all, management
ever, measuring biodiversity and its loss continues to pose would score zero. Where management had fully miti-
unique challenges across scales, particularly where there gated a threat, the score would be 100%. Thus, according
is a paucity of long-term data (Purvis & Hector 2000; to this scoring, threats to biodiversity could only either
Stedman-Edwards 2000; The Royal Society 2003; Wolman remain as they were or have positive mitigation. My jus-
2006). tification for modifying the TRA was that the original

Conservation Biology
Volume 22, No. 6, 2008
Anthony 1499

assessment design provides an oversimplistic and sessments of river-corridor health in South Africa show
potentially overoptimistic view of agencies’ abilities to that human activities outside protected areas can signifi-
mitigate threats. cantly affect the status of these ecosystems in protected
A second identified limitation concerns the standard- areas downstream (Foxcroft et al. 2007; Nel et al. 2007).
ization of threat types (Salafsky & Margoluis 1999). A South Africa’s dramatic policy changes in 1994 affected
number of attempts have been made to produce generic management of biodiversity within and outside protected
threat categories, with some overlap (EPA 1998; TNC areas (Cock & Fig 2000; Mabunda et al. 2003). In spite of
2000; Ervin 2003a; Salafsky et al. 2002; Sanderson et al. this benchmark year, no comprehensive baseline studies
2002). These attempts have been more formalized within were conducted at that time to assess threats to biodiver-
the now widely used “unified classification of direct sity in my study area, and there have been no subsequent
threats” (IUCN-CMP 2006), where direct threats are de- attempts to evaluate the effectiveness of policy changes
fined as “[t]he proximate (human) activities or processes collectively within KNP and DFED-EA administrations in
that have caused, are causing or may cause the destruc- mitigating threats to biodiversity. To address this I used
tion, degradation and/or impairment of biodiversity and TRAs with both the original and modified scoring meth-
natural processes.” For practitioners working with scarce ods to assess the effectiveness of KNP and DFED-EA man-
resources, standardization of threat types enhances cross- agement in mitigating threats to biodiversity along KNP’s
project learning and allows for greater application in the western border over a 10-year period, commencing in
comparison of temporal and spatial variation across sites 1994. I chose TRA methodology because it allows for
or even regions experiencing similar threats. Standardiz- comparison across sites (and institutions) and because
ing threat types can also facilitate transfer of successful it is a useful tool when little or no baseline data exist.
mitigation strategies between sites. Furthermore, an attempt was made to standardize identi-
fied threats into the IUCN-CMP classification to allow for
spatial and institutional comparison.
Case Study
South Africa is endowed with diverse natural resources Study Area
and is ranked as the third-most biodiverse country in The western border of KNP, from the Punda Maria gate
the world (DEAT 1998). Despite this richness, however, south to the Klein Letaba River (Fig. 1), has 3 of the 16
South Africa’s biodiversity is under threat due to habi- ecozones that exist within the KNP: Mopane/Bushwillow
tat loss, overexploitation, alien invasive species, and the Woodlands, Sandveld, and Riverine (Jacana Education
challenge of resolving conflicts between protected areas 2000). Kruger National Park has 8 main river catchments,
and local people (DEAT 1998; Perrings 2000; Steenkamp including the Shingwedzi and Letaba in the study area.
& Urh 2000; Anthony 2007). Annual precipitation ranges from 500 to 700 mm in the
Through provincial legislation and necessary coordi- area (Jacana Education 2000).
nating structures, environment and conservation depart- Land use adjacent to the western border of the KNP
ments in each province play an important role in devel- is characterized by slightly undulating plains containing
oping collaborations with other departments responsible villages surrounded by areas for subsistence farming. In
for activities concerning the conservation and use of bio- addition, there remain relatively sizeable vacant, bush-
diversity within the province. In Limpopo Province En- land areas, especially between the Shingwedzi and Klein
vironmental Affairs is a branch within the Department Letaba rivers (DWAF et al. 2001). Adjacent areas are de-
of Finance and Economic Development (DFED-EA). This marcated from the KNP with a boundary fence origi-
branch is operationally subdivided into municipal dis- nally intended to control the spread of foot-and-mouth
tricts, which provide conservation extension services and disease.
regulate and monitor the use of natural resources. The
DFED-EA’s activities are largely governed by the Limpopo
Environmental Management Act (LEMA) Number 7 of Methods
2003, which is aligned with national legislation. An addi-
tional function of DFED-EA is to promote sustainable de- I conducted 2 TRAs by organizing group discussions in
velopment outside protected areas through the forging of August 2004 with KNP representatives from the man-
appropriate partnerships with communities, NGOs, the agement and law enforcement departments from each
private sector, and other government departments. of the primary KNP ranger sections in the study area
Many authors have posited that conserving habitat in (Punda Maria, Shangoni). I conducted 2 additional TRAs
areas surrounding protected areas supports wildlife pop- with DFED-EA staff from the Greater Giyani and Mala-
ulations within it (Taylor 1982; Western & Gichohi 1993; mulele municipality offices. Participants either held po-
Homewood et al. 2001), and this is one of the reasons sitions within the study area at the time of the TRA or
driving efforts to maintain biodiversity outside Kruger had worked in the area for at least 10 years and were fa-
National Park (KNP) (Pollard et al. 2003). Recently, as- miliar with local biodiversity and its threats. To avoid

Conservation Biology
Volume 22, No. 6, 2008
1500 Threat Reduction Assessment

Figure 1. (a) Location of Kruger

National Park (KNP) in South
Africa and study area and (b)
areas for threat reduction
assessment (TRAs) (dotted lines)
and KNP border (bold line).

redundancy during discussions, I secondarily selected management had mitigated each threat. All participants
participants so the group would be diverse but would were given approximately 5 min to think about each
be able to communicate in the same language (Krueger threat and independently evaluate the extent to which
1994). management approaches had addressed a specific threat.
First, I asked KNP and DFED-EA staff to identify, in their Scores were assigned on a percentage basis. In contrast
own words, all threats to biodiversity, concentrating their to the original TRA scoring, the option for a negative
assessment on the period specified and area within their mitigation score was added for cases in which threats
jurisdiction. Jurisdictions were Punda Maria (KNP border had worsened, and a score of –100% was assigned to the
to 5 km inside park, from Luvuvhu River to the south threat if it arose after 1994 and had not been mitigated. Af-
end of ranger section); Shangoni (KNP border to 5 km ter the scoring and ranking exercise, total ranking scores
inside park, along entire western edge of ranger section); for each threat were multiplied by the percentage of the
Malamulele (KNP border fence to 15 km outside park, threat met to get a raw score for that threat. The threat
between the Luvuvhu and Shingwedzi rivers); and Giyani reduction index (TRA-I) value was derived by dividing
(KNP border fence to 15 km outside park, between the the sum of the raw scores for each threat by the total
Shingwedzi and Klein Letaba rivers). possible rankings of all the threats and multiplying by
Threats were defined as any human-related phenom- 100: (TRA-I =  total raw scores /  total rankings ×
ena that negatively affect biodiversity of the area in ques- 100) (Margoluis & Salafsky 2001). Where participants in-
tion (Margoluis & Salafsky 2001). Natural phenomena, dicated either a new or worsening threat, I inserted their
such as natural fires, were not considered threats, al- estimate of the degree to which the threat had worsened
though illegally lit fires were. into the TRA worksheet alongside the value (0) accord-
Second, to help participants focus their thinking about ing to the original TRA scoring design. This procedure
species richness, habitat condition and area, and ecosys- was carried out for the 2 ranger sections and 2 munici-
tem functioning, they were asked to rank threats accord- palities so a meaningful comparison of the indices inside
ing to their relative importance to one another. They con- and outside the park and throughout the study area could
sidered the portion of habitat(s) in the site that the threat be made. I compared original and modified TRA-I values
affects; the impact or severity of destruction caused by for the 4 areas with a paired t test, which allows compar-
the threat; and the urgency or immediacy of addressing isons of means between paired data (Wheater & Cook
the threat. A total sum score was computed after all the 2000). Threat definitions and what would constitute
threats were ranked. 100% mitigation for each threat across all assessment ar-
Third, a consensus building exercise was used with the eas are indicated in the TRA worksheets (see Supporting
groups to assess the extent to which the KNP or DFED-EA Information).

Conservation Biology
Volume 22, No. 6, 2008
Anthony 1501

Finally, an attempt was made to standardize threat Results

types by combining all threats for the 4 assessment areas
into general categories on the basis of IUCN-CMP classifi- The TRA-I values according to the original scoring ranged
cation (IUCN-CMP 2006). The standardized threats were from 16.6 (Punda Maria) to 50.0 (Shangoni), whereas val-
then prioritized on the basis of their presence or absence ues incorporating the modified scoring ranged from −5.4
in the 4 assessed areas. (Punda Maria) to 31.7 (Giyani) (Table 1 & Fig. 2). Modified

Table 1. Modified threat reduction assessment (TRA) results for Malamulele, Punda Maria, Giyani, and Shangoni assessment areas in South Africa.

Assessment Total Percent threat Raw

area Threat a Area Intensity Urgency rank reduced b scoreb

Malamulele illegal commercial harvesting of trees 10 7 10 27 −50 (0) −13.5 (0)

agricultural expansion 6 9 8 23 −40 (0) −9.2 (0)
illegal harvesting of trees for subsistence 7 6 9 22 −10 (0) −2.2 (0)
subsistence poaching 9 8 4 21 30 6.3
mining sand 2 10 5 17 40 6.8
illegal fire 4 5 7 16 60 9.6
residential expansion 5 4 6 15 −20 (0) −3 (0)
commercial poaching 8 2 3 13 40 5.2
road construction/maintenance 3 3 2 8 60 4.8
disease transfer 1 1 1 3 50 1.5
total 55 55 55 165 6.3 (34.2)
TRA index 3.8 (20.7)
Punda Maria poaching with dogs and/or snares 11 5 10 26 −15 (0) −3.9 (0)
poaching fish 5 11 8 24 −30 (0) −7.2 (0)
alien species 10 7 6 23 70 16.1
illegal harvesting of trees for medicine 2 10 11 23 −60 (0) −13.8 (0)
illegal fire 9 8 5 22 0 0
poaching with firearms 6 4 9 19 80 15.2
illegal harvesting of live 3 9 4 16 0 0
trees and/or dry wood
increasing elephant population 8 6 1 15 −60 (0) −9 (0)
highly infectious alien diseases 7 2 3 12 −80 (0) −9.6 (0)
commercial hunting–luring lions 1 3 7 11 5 0.55
endemic disease transfer 4 1 2 7 15 1.05
total 66 66 66 198 −10.6 (32.9)
TRA index −5.4 (16.6)
Giyani illegal harvesting of trees for subsistence 7 4 8 19 60 11.4
illegal fire 4 8 6 18 30 5.4
illegal commercial harvesting of trees 5 5 7 17 20 3.4
subsistence poaching 8 2 4 14 50 7
mining sand 2 7 5 14 −50 (0) −7 (0)
commercial poaching 6 3 3 12 40 4.8
road construction/maintenance 1 6 2 9 50 4.5
disease transfer 3 1 1 5 95 4.75
total 36 36 36 108 34.25 (41.25)
TRA index 31.7 (38.2)
Shangoni poaching wild animals 8 4 8 20 90 18
poaching fish 1 8 7 16 50 8
illegal fires 5 7 4 16 70 11.2
poaching grass/trees 7 2 5 14 50 7
commercial hunting 3 5 6 14 −100 (0) −14 (0)
increasing elephant population 4 6 2 12 −50 (0) −6 (0)
disease transfer 6 3 1 10 50 5
alien plant species 2 1 3 6 80 4.8
total 36 36 36 108 34 (54)
TRA index 31.5 (50.0)
a Detailed description of threats available (see Supporting Information).
b Values in parentheses incorporate original TRA scoring (see Methods).

Conservation Biology
Volume 22, No. 6, 2008
1502 Threat Reduction Assessment

Particularly problematic threats included the illegal har-

vesting of trees, some for medicinal purposes; alien in-
fectious diseases; agricultural expansion; poaching fish;
and an increasing KNP elephant population. Threat mit-
igation in southern assessment areas (Giyani, Shangoni)
was better, with TRA-I values of 31.7 and 31.5, respec-
tively, although participants in the Shangoni assessment
remarked that threats were more acute north of Shing-
wedzi River. Nevertheless, with 100 being an optimum
score, threats, such as illegal removal of sand, illegal com-
mercial hunting, and an increasing elephant population,
were getting worse.

From a methodological standpoint, the results raise a
number of noteworthy considerations. There is a need
to integrate negative values into the “% threat reduced”
Figure 2. Comparison of threat reduction assessment category of the original assessment design. By including
(TRA) index values for assessment areas with original negative values, where appropriate, investigators can get
and modified scoring. a more accurate picture of biodiversity threats and trends
temporally within assessment areas and spatially across
sites. In this study the TRA-I values of the original and
TRA-I values were significantly lower than values incor- modified versions were significantly different. In assess-
porating the original TRA scoring (t = 4.793, df = 3, p < ments with TRAs, incorporating negative scores may also
0.01). lead to marked differences in terms of prioritizing threats
Five of the 11 IUCN-CMP categorical threats were because new or worsening threats can be evaluated on
present in all 4 assessment areas, 2 were restricted to the basis of their degree of mitigation. Better prioritiza-
KNP, 2 were restricted to DFED-EA jurisdictions, and 2 tion of threats could lead to alternate policy choices in
were present only in Malamulele municipality (Table 2). response to identified threats.
Efforts in the northern sections of the study area (Mala- The interpretation of 100% threat mitigation also
mulele, Punda Maria) were less successful in mitigating presents a problem that is not easily rectified. If one as-
threats since 1994, with 11 of the 21 identified threats sumes 100% threat mitigation means some acceptable
in these areas showing no improvement with mitigation. level of threat, one is faced with a potentially awkward

Table 2. Combined prioritized ranking of standardized threats to biodiversity determined on the basis of presence or absence in assessment areas
(see IUCN-CMP 2006 for classifications).∗

Assessment area
no. Threat (IUCN-CMP classification) Malamulele Giyani Punda Maria Shangoni Total

5.1 hunting & collecting terrestrial animals X X X X 4

5.3 logging & wood harvesting X X X X 4
5.4 fishing & harvesting aquatic resources X X X X 4
7.1 fire & fire suppression X X X X 4
8.1 invasive non-native/alien species X X X X 4
5.2 gathering terrestrial plants X X 2
8.2 problematic native species (elephants) X X 2
3.2 mining & quarrying X X 2
4.1 roads & railroads X X 2
1.1 housing & urban areas X 1
2.1 annual & perennial nontimber crops X 1
total 9 7 7 7
∗ Abbreviations:
IUCN-CMP, World Conservation Union–Conservation Measures Partnership; DFED-EA, Department of Finance and Economic
Development–Environmental Affairs; KNP, Kruger National Park.

Conservation Biology
Volume 22, No. 6, 2008
Anthony 1503

use of percentage values. To illustrate, if one assumes where potential ineffective and counter-productive re-
100% mitigation means alien species are reduced from sponses may be made as a result of such generalizations.
30% cover to 10% cover in an assessment area, then miti- For example, if the goal of the assessment is to prior-
gation is successful by a TRA score of 50% if alien species itize proximate threats and thus to choose appropriate
cover is reduced by half (20% of the assessment area). conservation strategies, developing subcategories within
Nevertheless, if the threat worsens over the assessment broader threat classifications may be warranted. In do-
period to an extent that alien species spread to 70% of the ing so, more detailed threat descriptions can be retained.
area, then according to this scale, mitigation receives a Alternatively, assessments could be conducted in a step-
score of −200%, which is an awkward use of percentages. wise fashion considering scale (i.e., local or site, regional
To avoid overcomplicated scoring, I propose that practi- or national, and international), with codes at each level
tioners define 100% mitigation as the complete eradica- allowing threats to be summarized, depending on the
tion of a threat, irrespective of what initial condition or scale being investigated. In this way, detailed informa-
pressure the threat posed. Thus, in the above case, 100% tion is not lost, but can be generalized when necessary
mitigation of the alien species threat means 0% cover in to understand larger-scale threats and trends.
the assessment area. In this way, there can be no ambigu- Finally, although I propose modifications to TRA that
ity about what 100% mitigation means and, at the same should lead to a more accurate assessment of threats to
time, practitioners can still decide whether they are will- biodiversity, continued refinement of the approach is
ing to live with a certain percentage of threat. Moreover, needed. The following are particular avenues for con-
many demographic and socioeconomic indicators are re- tinuing efforts that would improve the usefulness of this
ported in absolute percentage values >100 (e.g., Baumert tool.
et al. 2005; Parsley & Halabisky 2008), especially when
making comparisons with a reference. Thus, it makes • Identify techniques to incorporate threat magnitude.
sense to adopt this terminology (rather than instituting a For example, if illegal mining of sand affected 1 river in
cap of –100% mitigation) when calculating the degree to an assessment site and this threat increased to 3 rivers
which a threat has worsened over an assessment period. over the assessment period, then the threat would have
If original TRA values have been used to measure con- increased by 200%, and mitigation would be given a
servation success, reevaluation of these cases may be TRA score of −200%. In such cases in which threats
valuable. This should be done if TRAs have been used have more than doubled, the original TRA scoring has
to develop strategies for resource management (Hast- an inherent weakness in that it cannot accurately mea-
ings & Fischer 2001) or when comparing effectiveness sure such changes. I propose defining 100% mitiga-
of conservation approaches (Mugisha & Jacobson 2004). tion as the complete elimination of a threat, but do
In such cases, results are likely to be less optimistic, with not suggest an upper limit to how bad a threat can
TRA-I values remaining where they are in best-case sce- potentially become (allowing for threats to increase
narios, but in all likelihood decreasing when incorporat- >100%). Because TRAs use relative rather than abso-
ing any new or worsening threats. lute values, how to standardize this aspect of threat
The standardization of threats continues to challenge magnitudes will need further development if cross-site
TRAs. I attempted to align categories with the IUCN-CMP comparisons are to be made.
direct threat classification on the basis of the nature of the • To improve application of the technique, TRAs should
threat, akin to content coding in qualitative data analysis be conducted at the onset and over multiple time pe-
(Taylor & Bogdan 1984; Weisberg et al. 1996). On one riods of a conservation intervention to identify trends
hand, this categorization has helped KNP and DFED-EA in specified threats. In my study, no such baseline as-
staff conceptualize local threats to biodiversity and may sessment was made, and results would have benefited
lead to more focused action and appropriate response. from previously conducted TRAs. Longitudinal studies
On the other hand, generalizing threats, similar to cod- utilizing TRAs would improve the adaptive manage-
ing qualitative data, may inevitably result in a loss of rich- ment utility of the approach and allow practitioners
ness of the threat description, leading to responses that to estimate probable time frames for reaching 100%
may not necessarily address the locally identified threat. abatement.
For example, subsistence poaching, poaching with dogs • Finally, compiling a database of successful strategies
or snares, poaching with firearms, and luring of lions to combat specific threats, which can be easily ac-
for illegal commercial hunting are all combined into the cessed by managers, practitioners, and researchers,
“hunting & collecting terrestrial animals” (IUCN-CMP no. would be a useful secondary outcome of the TRA
5.1) categorical threat in this study. Nevertheless, strate- methodology. This may be integrated into ongoing ini-
gies to combat each of these threats may require quite tiatives including the IUCN and CMP’s Classification of
discrete policies and actions, drawing on different legis- Conservation Actions [
lation and institutions. Careful attention must be taken ssc/sis/classification.htm] or Conservation Evidence
to understand the reasons for threat standardization and [].

Conservation Biology
Volume 22, No. 6, 2008
1504 Threat Reduction Assessment

Acknowledgments Hockings, M., S. Stolton, and N. Dudley. 2000. Evaluating effective-

ness: a framework for assessing the management of protected areas.
IUCN, Gland, Switzerland.
I thank the CEU Doctoral Research Support Program for Homewood, K., E. F. Lambin, E. Coast, A. Kariuki, I. Kikula, J. Kivelia,
funding, staff from Conservation Services and People & M. Said, S. Serneels, and M. Thompson. 2001. Long-term changes in
Conservation (KNP), and Mopani and Vhembe districts Serengeti–Mara wildebeest and land cover: pastoralism, population
(DFED-EA) for support. I also thank 3 anonymous review- or policies? Proceedings of the National Academy of Sciences of the
ers and the assigning editor for constructive comments United States of America 98:12544–12549.
IUCN-CMP (International Union for the Conservation of Nature-
on an earlier version of this manuscript. Conservation Measures Partnership). 2006. Unified classifica-
tion of direct threats. Version 1.0. Available from http://www. (accessed April 2008).
Supporting Information Jacana Education. 2000. Kruger National Park ecozone map. Jacana
Education, Johannesburg.
The TRA worksheets for Malamulele (Appendix S1), Krueger, R. A. 1994. Focus groups: a practical guide for ap-
Punda Maria (Appendix S2), Giyani (Appendix S3), and plied research. 2nd edition. Sage Publications, Thousand Oaks,
Shangoni (Appendix S4) are available as part of the on- Mabunda, D., D. J. Pienaar, and J. Verhoef. 2003. The Kruger National
line article. The author is responsible for the content Park: a century of management and research. Pages 3–21 in J. T. du
and functionality of these materials. Queries (other than Toit, K. H. Rogers, and H. Biggs, editors. The Kruger experience:
absence of the material) should be directed to the corre- ecology and management of savanna heterogeneity. Island Press,
sponding author. Washington, D.C.
Margoluis, R., and N. Salafsky. 1998. Measures of success: designing,
managing, and monitoring conservation and development projects.
Literature Cited Island Press, Washington, D.C.
Margoluis, R., and N. Salafsky. 2001. Is our project succeeding? A guide
Anthony, B. 2007. The dual nature of parks: attitudes of neighbouring to threat reduction assessment for conservation. Biodiversity Sup-
communities towards Kruger National Park, South Africa. Environ- port Program, Washington, D.C.
mental Conservation 34:236–245. MEA (Millennium Ecosystem Assessment). 2005. Ecosystems and hu-
Baumert, K. A., T. Herzog, and J. Pershing. 2005. Navigating the num- man well-being: biodiversity synthesis. World Resources Institute,
bers: greenhouse gas data and international climate policy. World Washington, D.C.
Resources Institute, Washington, D.C. Mugisha, A. R., and S. K. Jacobson. 2004. Threat reduction assessment
CMP (Conservation Measures Partnership). 2004. Open standards for of conventional and community-based conservation approaches to
the practice of conservation. Version 1. Conservation Measures Part- managing protected areas in Uganda. Environmental Conservation
nership, Washington, D.C. 31:233–241.
Cock, J., and D. Fig. 2000. From colonial to community based conser- Nel, J. L., D. J. Roux, G. Maree, C. J. Kleynhans, J. Moolman, B. Reyers, M.
vation: environmental justice and the national parks in South Africa. Rouget, and R. M. Cowling. 2007. Rivers in peril inside and outside
Society in Transition 31:22–35. protected areas: a systematic approach to conservation assessment
Costanza, R., et al. 1997. The value of the world’s ecosystem services of river ecosystems. Diversity and Distributions 13:341–352.
and natural capital. Nature 387:253–260. Noss, R. F. 1990. Indicators for monitoring biodiversity: a hierarchical
DEAT (Department of Environmental Affairs and Tourism). 1998. Con- approach. Conservation Biology 4:355–364.
vention on biological diversity: South African national report to the Parsley, C., and D. Halabisky. 2008. Profile of growth firms: a
Fourth Conference of the Parties. Department of Environmental Af- summary of Industry Canada research. Industry Canada, Ot-
fairs and Tourism, Pretoria. tawa. Available from
DWAF (Department of Water Affairs and Forestry), DEAT (Depart- en/h_rd01200e.html (accessed April 2008).
ment of Environmental Affairs and Tourism), and WRC (Water Perrings, C. 2000. The biodiversity convention and biodiversity loss in
Research Commission). 2001. State of the rivers report: Letaba Sub-Saharan Africa. Pages 1–43 in C. Perrings, editor. The economics
and Luvuvhu River systems. WRC report TT 165/01. Water Re- of biodiversity conservation in Sub-Saharan Africa: mending the Ark.
search Commission, Pretoria. Available from Edward Elgar, Cheltenham, United Kingdom.
rhp/state_of_rivers/letluv_01_toc.html (accessed June 2007). Persha, L. 2001. Threat reduction assessment (TRA) notes and method-
Ehrlich, P. R., and A. H. Ehrlich. 1992. The value of biodiversity. Ambio ology changes. UNDP-GEF East Africa Cross Borders Biodiversity
21:219–226. Project, Arusha, Tanzania.
EPA (Environmental Protection Agency). 1998. Guidelines for ecologi- Pollard, S., C. M. Shackleton, and J. Carruthers. 2003. Beyond the fence:
cal risk assessment. EPA/630/R-95/002F. Environmental Protection people and the lowveld landscape. Pages 422–446 in J. T. du Toit,
Agency, Washington, D.C. K. H. Rogers, and H. C. Biggs, editors. The Kruger experience:
Ervin, J. 2003a. WWF: rapid assessment and prioritization of protected ecology and management of savanna heterogeneity. Island Press,
area management (RAPPAM) methodology. World Wide Fund for Washington, D.C.
Nature, Gland, Switzerland. Purvis, A., and A. Hector. 2000. Getting the measure of biodiversity.
Ervin, J. 2003b. Rapid assessment of protected area manage- Nature 405:212–219.
ment effectiveness in four countries. BioScience 53:833– Reaka-Kudla, M. L., D. E. Wilson, and E. O. Wilson, editors. 1997. Bio-
841. diversity II: understanding and protecting our biological resources.
Foxcroft, L. C., M. Rouget, and D. M. Richardson. 2007. Risk assess- Joseph Henry Press, Washington, D.C.
ment of riparian plant invasions into protected areas. Conservation Salafsky, N., and R. Margoluis. 1999. Threat reduction assessment: a
Biology 21:412–421. practical and cost-effective approach to evaluating conservation and
Hastings, R. M., and D. W. Fischer. 2001. Management priorities for Mag- development projects. Conservation Biology 13:830–841.
dalena Bay, Baja California, Mexico. Journal of Coastal Development Salafsky, N., R. Margoluis, K. H. Redford, and J. G. Robinson. 2002. Im-
7:193–202. proving the practice of conservation: a conceptual framework and

Conservation Biology
Volume 22, No. 6, 2008
Anthony 1505

research agenda for conservation. Conservation Biology 16:1469– The Royal Society. 2003. Measuring biodiversity for conservation. Doc-
1479. ument 11/03. The Royal Society, London.
Sanderson, E. W., K. H. Redford, A. Vedder, P. B. Coppolillo, and S. E. TNC (The Nature Conservancy). 2000. The five-S framework for site
Ward. 2002. A conceptual model for conservation planning based conservation: a practitioner’s handbook for site conservation plan-
on landscape species requirements. Landscape and Urban Planning ning and measuring conservation success. The Nature Conservancy,
58:41–56. Arlington, Virginia.
Stedman-Edwards, P. 2000. A framework for analyzing biodiversity loss. TNC (The Nature Conservancy). 2003. The enhanced five-S project
Pages 11–35 in A. Wood, P. Stedman-Edwards, and J. Mang, editors. management process: an overview of proposed standards for de-
The root causes of biodiversity loss. Earthscan, London. veloping strategies, taking action, and measuring effectiveness
Steenkamp, C., and J. Urh. 2000. Discovering power relations in a South and status at any scale. The Nature Conservancy, Arlington,
African CBNRM case study: the Makuleke community of the North- Virginia.
ern Province. Pages 126–155 in S. Shackleton and B. Campbell, ed- Tucker, G. 2005. A review of biodiversity conservation performance
itors. Empowering communities to manage natural resources: case measures. Earthwatch Institute, Oxford, United Kingdom.
studies from Southern Africa. CSIR, Pretoria, South Africa. WCS (Wildlife Conservation Society). 2002. Using conceptual models
Stem, C., R. Margoluis, N. Salafsky, and M. Brown. 2005. Monitoring to set conservation priorities. Living Landscapes Bulletin 5:1–4.
and evaluation in conservation: a review of trends and approaches. Weisberg, H. F., J. A. Krosnick, and B. D. Bowen. 1996. An introduc-
Conservation Biology 19:295–309. tion to survey research, polling, and data analysis. 3rd edition. Sage
Taylor, R. D. 1982. Buffer zones: resolving conflict between human and Publications, Thousand Oaks, California.
wildlife interests in the Sebungwe region. Zimbabwe Agricultural Western, D., and H. Gichohi. 1993. Segregation effects and the impov-
Journal 79:179–184. erishment of savanna parks: the case for ecosystem viability analysis.
Taylor, S. J., and R. Bogdan. 1984. Introduction to qualitative research African Journal of Ecology 31:268–271.
methods: the search for meanings. Wiley, New York. Wheater, C. P., and P. A. Cook. 2000. Using statistics to understand the
Teder, T., M. Moora, E. Roosaluste, K. Zobel, M. Partel, U. Koljalg, and M. environment. Routledge, New York.
Zobel. 2007. Monitoring of biological diversity: a common-ground Wolman, A. G. 2006. Measurement and meaningfulness in conservation
approach. Conservation Biology 21:313–317. science. Conservation Biology 20:1626–1634.

Conservation Biology
Volume 22, No. 6, 2008