ArticlePDF Available

Expansion of Industrial Logging in Central Africa

DOI:10.1126/science.1141057
Authors:
Nadine Laporte at Northern Arizona University
Nadine Laporte
Jared Stabach at Smithsonian Institution
Jared Stabach
Robert Grosch
Robert Grosch
Robert Grosch
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Tiffany S Lin
Tiffany S Lin
Tiffany S Lin
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Show all 5 authorsHide
Download full-text PDF
Read full-text
Download citation

Abstract and Figures

Industrial logging has become the most extensive land use in Central Africa, with more than 600,000 square kilometers (30%) of forest currently under concession. With use of a time series of satellite imagery for the period from 1976 to 2003, we measured 51,916 kilometers of new logging roads. The density of roads across the forested region was 0.03 kilometer per square kilometer, but areas of Gabon and Equatorial Guinea had values over 0.09 kilometer per square kilometer. A new frontier of logging expansion was identified within the Democratic Republic of Congo, which contains 63% of the remaining forest of the region. Tree felling and skid trails increased disturbance in selectively logged areas.
Figures - uploaded by Nadine Laporte
Author content
All figure content in this area was uploaded by Nadine Laporte
Content may be subject to copyright.
Content uploaded by Scott J Goetz
Author content
All content in this area was uploaded by Scott J Goetz
Content may be subject to copyright.
Expansion of Industrial Logging
in Central Africa
Nadine T. Laporte,* Jared A. Stabach, Robert Grosch, Tiffany S. Lin, Scott J. Goetz
C
entral Africa’s dense humid forests have
long been regarded as among the most
pristine on Earth, but in recent decades
industrial logging has become the most extensive
form of land use in the region. Currently more
than 600,000 km
2
(30%) of forest are under
logging concessions, whereas just 12% is pro-
tected. Logging-related disturbance in the region
alters ecosystem composition and biodiversity (1),
opens remote areas to poaching (2), and modifies
numerous functional attributes of the ecosystem
(3). Laws and regulations are in place to improve
forest management at national and regional scales,
but limited resources are available to enforce
regulations or to provide technical support (4).
Here, we report on the use of over 300 Landsat
satellite i mages, coveri ng 4 million km
2
, to track
the progression of logging roads for three decades
preceding 2003 (5). W e document accelerating
rates of logging road construction in much of the
region and show that monitoring with satellite
remote sensing provides a practical approach to
map changes associated with logging activities.
We mapped 51,916 km of logging roads with-
in the forested region (Fig. 1). This is a conserv-
ative estimate, because not all areas had recent
cloud-free satellite images (5) and logging roads
are converted to public roads where population
density is high. Logging roads accounted for 38%
of the length of all roads, ranging from 13% in the
Democratic Republic of Congo (DRC) to >60% in
Gabon and the Republic of Congo (ROC). The
combined road density (public and logging) was
0.07 km km
–2
and, considering logging roads only,
0.03 km km
–2
. The highest logging road densities
(0.09 km km
–2
) were in Cameroon and Equatorial
Guinea (EG), where most of the forest was cut at
least once. Logging in these two countries and in
Gabon has extended inland in recent decades after
the earlier harvesting of coastal forest.
The most rapidly changing area was in north-
ern ROC, where the rate of road construction in-
creased from 156 km year
−1
for the period 1976–1990
to over 660 km year
−1
after 2000. Evidence for a
new frontier of logging expansion was documented
within the DRC, which currently contains 63% of
the total remaining forest of the region and has the
lowest measured logging road density (0.01 km
km
–2
) of all Central African nations. Rates of
logging road construction increased within DRC,
particularly in a 50,000-km
2
region of north-central
DRC, where the rate of road development pro-
gressed from 336 km year
−1
(1986–1990) to 456 km
year
−1
(2000–20 0 2 ) . We expect industrial logging
concessions to expand, with commensurate in-
creases in the rates of logging and road building
associated with foreign investment (6).
W ith the exception of the Okoumé forests of
Gabon (7), most of the industrial logging is selec-
tive and focused on high-value tree species for
export (for example, African mahoganies). We esti-
mated 5% (89,715 km
2
)ofthetotalforestedareaas
disturbed and 29% (567,782 km
2
)asmorelikelyto
have increased wildlife hunting pressure because of
easier access and local market opportunities offered
by new loggin g towns (5). The greatest amount of
forest disturbance (15%) occurred in Cameroon
and EG, compared with just 1% within the DRC. In
addition, we used finer resolution (4-m) satellite
imagery to document disturbance created by log-
ging skid trails and tree felling. These locally dis-
turbed areas had canopy gaps that were five to six
times larger than those in adjacent unlogged forests.
Gaps created by tree fall ranged from 200 to 600 m
2
in size and, together with skid trails, accounted for
9% of the ar ea in which logging occurred .
To date, few reliable data sets have been avail-
able to monitor the changes taking place in remote
areas of the Congo Basin, but regular monitoring
with satellite remote sensing provides a consistent
approach to monitor both legal and illegal logging
activities. In the context of the rapid frontier ex-
pansion, the conservation of forested landscapes
and sustainable timber production is crucial for
Central African nations and their inhabitants.
References and Notes
1. J. R. Malcolm, J. C. Ray, Conserv. Biol. 14, 1623 (2000).
2. J. G. Robinson, K. H. Redford, E. L. Bennett, Science 284,
595 (1999).
3. J. S. Hall, D. J. Harris, V. Medjibe, P. M. S. Ashton,
For. Ecol. Manage. 183, 249 (2003).
4. The Central Africa Forests Commission, www.comifac.org.
5. Materials and methods are available on Science Online.
6. P. Buys, U. Deichmann, D. Wheeler, “Road network
upgrading and overland trade expansion in sub-Saharan
Africa” (Policy Research Working Paper WPS 4097, World
Bank, Washington, DC, 2006); available online at
www.worldbank.org/reference.
7. L. J. T. White, J. Trop. Ecol. 10, 309 (1994).
8. Supported by NASA grants NNG05GD14G and
NNS06AA06A. We thank E. Davidson, R. A. Houghton,
and D. Nepstad for comments on the manuscript.
Supporting Online Material
www.sciencemag.org/cgi/content/full/316/5830/1451/DC1
Materials and Methods
Fig. S1
Table S1
References and Notes
8 February 2007; accepted 26 April 2007
10.1126/science.1141057
BREVIA
Woods Hole Research Center, Falmouth, MA 02540, USA.
*To whom correspondence should be addressed. E-mail:
nlaporte@whrc.org
Fig. 1. Logging concessions and road distribution in Central Africa: Cameroon (labeled on the map as 1),
Central African Republic (2), Equatorial Guinea (3), Gabon (4), Republic of Congo (5), and Democratic Republic
of Congo (6). A more detailed graphic of logging roads for a portion of the region is provided in fig. S1.
www.sciencemag.org SCIENCE VOL 316 8 JUNE 2007
1451
on June 8, 2007 www.sciencemag.orgDownloaded from
... The environment around the village Djoube has been described as Congolian Lowland Forests (Loubelo Madiela, 2018;Jang and Boyette, 2021), often with clear signs of logging activity. This logging activity causes rapid deforestation in those areas where the BaYaka are residing (Lewis and Nelson, 2006;Laporte et al., 2007). At each camp site, rainfall and temperature data were recorded daily over multiple fieldwork periods from 2015 to 2020. ...
... Compared to adults, with the men hunting at night, children might focus more on easily collected plant foods than on game (Crittenden et al., 2013). The apparent reduced meat consumption could, on the other hand, be explained by the decreased population sizes of animal species and the increased hunting pressure coinciding with the logging activities (Yasuoka, 2006;Laporte et al., 2007). This logging activity was especially visible around camp Bongo. ...
Development of embodied capital: Diet composition, foraging skills, and botanical knowledge of forager children in the Congo Basin
Article
Full-text available
  • Feb 2023
The embodied capital theory states that the extended juvenile period has enabled human foragers to acquire the complex foraging skills and knowledge needed to obtain food. Yet we lack detailed data on how forager children develop these skills and knowledge. Here, we examine the seasonal diet composition, foraging behavior, and botanical knowledge of Mbendjele BaYaka forager children in the Republic of the Congo. Our data, acquired through long-term observations involving full-day focal follows, show a high level of seasonal fluctuation in diet and foraging activities of BaYaka children, in response to the seasonal availability of their food sources. BaYaka children foraged more than half of the time independent from adults, predominantly collecting and eating fruits, tubers, and seeds. For these most-consumed food types, we found an early onset of specialization of foraging skills in children, similar to the gendered division in foraging in adults. Specifically, children were more likely to eat fruit and seed species when there were more boys and men in the group, and girls were more likely than boys to collect tuber species. In a botanical knowledge test, children were more accurate at identifying plant food species with increasing age, and they used fruits and trunks for species identification, more so than using leaves and barks. These results show how the foraging activities of BaYaka children may facilitate the acquisition of foraging skills and botanical knowledge and provide insights into the development of embodied capital. Additionally, BaYaka children consumed agricultural foods more than forest foods, probably reflecting BaYaka’s transition into a horticultural lifestyle. This change in diet composition may have significant consequences for the cognitive development of BaYaka children.
View
... Tree removal opens the forest canopy, damages adjacent vegetation, and increases the penetration of solar radiation into the forest (Edwards et al., 2014). The effects of forest modification are early associated with the simplification in tree structure (Putz et al., 2012), although in successional stages after disturbance, certain areas may be severely affected (e.g., loss of regeneration) or colonized by heliophyte species related to edges and secondary forests, thus modifying both the horizontal and vertical structure of the forest (Hubbell et al., 1999;Laporte et al., 2007;Wagner et al., 2011;Hawthorne et al., 2012;Názaro et al., 2021). Despite the existence of substantial evidence on the impact of logging on certain structural attributes, the effect of disturbance on functional attributes has been scarcely quantified, especially in neotropical dry forests. ...
Post-logging changes in a neotropical dry forest composition and structure modify the ecosystem functioning
Article
Full-text available
Native forests are considered to be the main centers of terrestrial diversity in the world and they provide important ecosystem services for humanity. Unsustainable logging is currently-one of the greatest threats to the biodiversity of these forests. In northwestern Argentina, about 90 % of the piedmont forest in the Southern Yungas has been transformed and the remaining 10% is being intensively logged. The aim of this study was to identify post-logging changes in the structural and compositional attributes affecting the functionality of a dry forest in Argentina. In the control forest, we recorded 1559 individuals from 45 tree species and 22 families, and at post-logging forest, 1376 trees from 47 species and 25 families. Logging affected the forest structure by decreasing timber tree density and basal area. In addition, at the post-logging forest, we observed an increase in the density of non-timber trees and in the horizontal understory cover, which were associated with the normalized difference vegetation index, thus, affecting the forest functioning. The threatened timber species Amburana cearensis and Jacaranda mimosifolia were not recorded in the logged forest. In the control, four timber species had the highest importance value indexes (IVI): Phyllostilon rhamnoides, Anadenanthera colubrina, Calicophylum multiflorum, and Myracrodruon urundeuva. Additionally, in the logged forest there was a change in the species order regarding the IVI, with A. colubrina, P. rhamnoides, Cordia americana, and a non-valuable timber species (Ocotea puberula) being the most important tree species. In order to preserve the economic and ecological value of the piedmont forest, the density of non-timber tree species and the horizontal understory cover must be maintained at values close to the control forest. These guidelines can be implemented by applying low-impact logging techniques.
View
... The density of primary logging roads (0.064 km/km 2 ) was lower, around one seventh, than what we had seen previously in Borneo (using similar techniques, Gaveau et al. (2014), found 0.48 km/km 2 ) and comparable to the densities observed in Central Africa during the 1980s and 1990s (Laporte et al., 2007). Nonetheless, extensive forest cover means that most developments impact these unique ecosystems. ...
View
... Hosonuma et al. (2012) observed that over 70% of the direct drivers of deforestation are commercial and subsistence agriculture, whereas forest degradation is mainly due to logging and fuelwood extraction. Others argue that logging primarily causes deforestation (Laporte et al., 2007), and population growth (Jha and Bawa, 2006;DeFries et al., 2010), income and literacy rate , subsistence farming, and the extraction of fuelwood (Brink and Eva, 2009;Fisher, 2010;Boucher et al., 2011;Sanford et al., 2011) drive deforestation. Weatherley-Singh andGupta (2015) observed that agriculture conversion, logging, fuelwood extraction, roads and infrastructure, mining, forest fires, land tenure and ownership, and poverty are drivers of deforestation and forest degradation. ...
Towards REDD+ implementation: Deforestation and forest degradation drivers, REDD+ financing, and readiness activities in participant countries
Article
Full-text available
  • Oct 2022
This study provides an assessment of dominant drivers of deforestation and forest degradation across 29 REDD+ participant countries and total funds disbursed from the Forest Carbon Partnership Facility to 15 participant countries since they signed their respective Readiness Preparation Proposals (R-PPs), as well as what they spent the grants on. This study aims to provide information for policymakers and donors on the need to invest in REDD+, as well as the need for participant countries to channel funds to key activities for effectively and efficiently implementing REDD+. The assessment is based on participant countries’ R-PPs and FCPF readiness fund progress reports, using a content analysis approach. The expansion of agriculture, which includes both commercial and subsistence agriculture, was the predominant driver of deforestation, while logging, such as industrial and small-scale exploitation, was the main driver of forest degradation in most participating countries. In addition, we observed that the 15 participant countries received a total of USD 702.36 million for REDD+ readiness preparation. However, most countries have received less than USD 10 million for their readiness-related activities. We observed a huge disparity regarding the funds received between countries from sub-Saharan Africa and their Asian and Latin American counterparts, particularly Indonesia and Peru, respectively. Furthermore, almost all countries have spent their funds on policy and strategy development, stakeholder and consultation events, and public awareness creation. Relatively small amounts have been spent on capacity building and training. Thus, international donors need to focus on strengthening institutional capacities and building effective forest governance structures in countries participating in REDD+, and they should also provide platforms to engage with participant countries to tackle the main deforestation and forest degradation drivers. We suggest further studies to identify high-performing countries that have received lower-than-average funding, consider their deforestation pressures, the size of forests, and the emission reduction potential.
View
Grassland cover declined in Southern Africa but increased in other African subcontinents in early 21st century
Preprint
Full-text available
  • Sep 2022
The African continent has the most extensive grassland cover in the world, providing valuable ecosystem services. African grasslands, like other continental grasslands, are prone to various anthropogenic disturbances and climate, and require data-driven monitoring for efficient functioning and service delivery. Yet, knowledge of how the African grassland cover has changed in the past years is lacking, especially at the subcontinents level, due to a lack of relevant long-term, Africa-wide observations and experiments. Here, we used Moderate Resolution Imaging Spectroradiometer (MODIS) Land Cover Type (MCD12Q1) data spanning 2001 to 2017 to conduct Land Use Land Cover (LULC) change analyses and map grassland distribution in Africa. Specifically, we assessed the changes in grassland cover across and within African subcontinents over three periods (2001–2013, 2013–2017, and 2001–2017). We found that the African grassland cover was 16777765.5 km ² , 16999468.25 km ² , and 16968304.25 km ² in 2001, 2013, and 2017, respectively. There were net gain (1.32%) and net loss (0.19%) during 2001–2013 and 2013–2017 periods, respectively. Generally, the African grassland cover increased (1.14%) over the whole study period (2001–2017) at the expense of forestland, cropland, and built-up areas. The East and West African grassland covers reduced by 0.07% and 1.35%, respectively, in 2013–2017, and increased in other periods; those of North and Central Africa increased throughout the three periods; but Southern Africa grassland cover decreased over the three periods. Overall, the net gains in the grassland cover of other African subcontinents offset the loss in Southern Africa and promoted the overall gain across Africa. This study underscores the need for continuous monitoring of African grasslands and the causes of their changes for efficient delivery of ecosystem services.
View
Carbon and nutrient fluxes across tropical river-coastal boundaries
Chapter
  • Oct 2013
This volume provides a state-of-the-art summary of biogeochemical dynamics at major river-coastal interfaces for advanced students and researchers. River systems play an important role (via the carbon cycle) in the natural self-regulation of Earth's surface conditions by serving as a major sink for anthropogenic CO2. Approximately 90 percent of global carbon burial occurs in ocean margins, with the majority of this thought to be buried in large delta-front estuaries (LDEs). This book provides information on how humans have altered carbon cycling, sediment dynamics, CO2 budgets, wetland dynamics, and nutrients and trace element cycling at the land-margin interface. Many of the globally important LDEs are discussed across a range of latitudes, elevation and climate in the drainage basin, coastal oceanographic setting, and nature and degree of human alteration. It is this breadth of examination that provides the reader with a comprehensive understanding of the overarching controls on major river biogeochemistry.
View
Flux and fate of the Yellow (Huanghe) River–derived materials to the sea: impacts of climate change and human activities
Chapter
  • Oct 2013
This volume provides a state-of-the-art summary of biogeochemical dynamics at major river-coastal interfaces for advanced students and researchers. River systems play an important role (via the carbon cycle) in the natural self-regulation of Earth's surface conditions by serving as a major sink for anthropogenic CO2. Approximately 90 percent of global carbon burial occurs in ocean margins, with the majority of this thought to be buried in large delta-front estuaries (LDEs). This book provides information on how humans have altered carbon cycling, sediment dynamics, CO2 budgets, wetland dynamics, and nutrients and trace element cycling at the land-margin interface. Many of the globally important LDEs are discussed across a range of latitudes, elevation and climate in the drainage basin, coastal oceanographic setting, and nature and degree of human alteration. It is this breadth of examination that provides the reader with a comprehensive understanding of the overarching controls on major river biogeochemistry.
View
The Nile delta in the anthropocene: drivers of coastal change and impacts on land-ocean material transfer
Chapter
  • Oct 2013
This volume provides a state-of-the-art summary of biogeochemical dynamics at major river-coastal interfaces for advanced students and researchers. River systems play an important role (via the carbon cycle) in the natural self-regulation of Earth's surface conditions by serving as a major sink for anthropogenic CO2. Approximately 90 percent of global carbon burial occurs in ocean margins, with the majority of this thought to be buried in large delta-front estuaries (LDEs). This book provides information on how humans have altered carbon cycling, sediment dynamics, CO2 budgets, wetland dynamics, and nutrients and trace element cycling at the land-margin interface. Many of the globally important LDEs are discussed across a range of latitudes, elevation and climate in the drainage basin, coastal oceanographic setting, and nature and degree of human alteration. It is this breadth of examination that provides the reader with a comprehensive understanding of the overarching controls on major river biogeochemistry.
View
Changjiang (Yangtze) and Huanghe (Yellow) Rivers: historical reconstruction of land-use change and sediment load to the sea
Chapter
  • Oct 2013
This volume provides a state-of-the-art summary of biogeochemical dynamics at major river-coastal interfaces for advanced students and researchers. River systems play an important role (via the carbon cycle) in the natural self-regulation of Earth's surface conditions by serving as a major sink for anthropogenic CO2. Approximately 90 percent of global carbon burial occurs in ocean margins, with the majority of this thought to be buried in large delta-front estuaries (LDEs). This book provides information on how humans have altered carbon cycling, sediment dynamics, CO2 budgets, wetland dynamics, and nutrients and trace element cycling at the land-margin interface. Many of the globally important LDEs are discussed across a range of latitudes, elevation and climate in the drainage basin, coastal oceanographic setting, and nature and degree of human alteration. It is this breadth of examination that provides the reader with a comprehensive understanding of the overarching controls on major river biogeochemistry.
View
Anthropogenicchanges in sediment and nutrient retention in the Rhine delta
Chapter
  • Oct 2013
This volume provides a state-of-the-art summary of biogeochemical dynamics at major river-coastal interfaces for advanced students and researchers. River systems play an important role (via the carbon cycle) in the natural self-regulation of Earth's surface conditions by serving as a major sink for anthropogenic CO2. Approximately 90 percent of global carbon burial occurs in ocean margins, with the majority of this thought to be buried in large delta-front estuaries (LDEs). This book provides information on how humans have altered carbon cycling, sediment dynamics, CO2 budgets, wetland dynamics, and nutrients and trace element cycling at the land-margin interface. Many of the globally important LDEs are discussed across a range of latitudes, elevation and climate in the drainage basin, coastal oceanographic setting, and nature and degree of human alteration. It is this breadth of examination that provides the reader with a comprehensive understanding of the overarching controls on major river biogeochemistry.
View
Wildlife Harvest in Logged Tropical Forests
Article
Full-text available
Commercial logging, by opening up remote tropical forest areas, stimulating a commercial trade, and bringing in people from other areas, has hugely increased the harvest of wildlife. This loss of wildlife threatens the ability of forest-dwelling people to feed themselves, the survival of harvested species, and the sustainability of tropical forestry itself. As commercial forestry has created the conditions for the increased wildlife harvest, regulatory mechanisms should focus on forestry operations. While progress has been slight to date, there is some openness within the industry to the idea of sustaining all elements of the rain forest ecosystem, including wildlife, when forests are logged.
View
The effects of selective logging on forest structure and tree species composition in a Central African forest: Implications for management of conservation areas
Article
The forests of Central Africa enjoy world-wide recognition for their spectacular wildlife and also harbor an abundance of high quality timber. With mismanagement and the conversion of large tracts of West African forest to agricultural production, Central African forests are experiencing increased harvesting pressures. This is particularly true for species of African mahogany (Entandrophragma spp.).In the tri-national region of Cameroon, Central African Republic, and Republic of Congo, a widely applied version of the Integrated Conservation and Development Model attributes the dual management objectives of biodiversity conservation and timber production to the same zones. Many conservationists working in the region believe that highly selective timber extraction is the best management scenario to meet these objectives. Conventional wisdom holds that if selective logging does not adequately regenerate Entandrophragma spp., loggers will quit the region after having mined the forest.A comparison between unlogged, 6-month and 18-year post-harvest forest stands indicates lasting effects of highly selective, high grade logging. While there was little difference in tree species composition and diversity between treatments, stem densities of both saplings and trees in unlogged forest were significantly higher than those in forest sampled 18 years after logging. Evidence suggests inadequate recruitment of Entandrophragma cylindricum and E. utile, the principal timber species, to justify continued timber extraction. Data indicate a significant shift in canopy dominance from shade intolerant to shade bearing species due to insufficient canopy disturbance. Nevertheless, an abundance of other top quality timber species remains after selective removal of African mahogany and these forests will remain attractive to loggers long after the elimination of Entandrophragma spp. A better approach to manage timber zones for timber production and conservation would be an adaptive management approach based on increased species selection and canopy disturbance. Zones targeting the conservation of closed forest obligate species should not be logged.
View
Road network upgrading and overland trade expansion in Sub-Saharan Africa
Article
  • Jan 2007
Recent research suggests that isolation from regional and international markets has contributed significantly to poverty in many Sub-Saharan African countries. Numerous empirical studies identify poor transport infrastructure and border restrictions as significant deterrents to trade expansion. In response, the African Development Bank has proposed an integrated network of functional roads for the subcontinent. Drawing on new econometric results, the authors quantify the trade-expansion potential and costs of such a network. They use spatial network analysis techniques to identify a network of primary roads connecting allSub-Saharan capitals and other cities with populations over 500,000. The authors estimate current overland trade flows in the network using econometrically-estimated gravity model parameters, road transport quality indicators, actual road distances, and estimates of economic scale for cities in the network. Then they simulate the effect of feasible continental upgrading by setting network transport quality at a level that is functional, but less highly developed than existing roads in countries like South Africa and Botswana. The authors assess the costs of upgrading with econometric evidence from a large World Bank database of road project costs in Africa. Using a standard approach to forecast error estimation, they derive a range of potential benefits and costs. Their baseline results indicate that continental network upgrading would expand overland trade by about $250 billion over 15 years, with major direct and indirect benefits for the rural poor. Financing the program would require about $20 billion for initial upgrading and $1 billion annually for maintenance. The authors conclude with a discussion of supporting institutional arrangements and the potential cost of implementing them.
View
  • Jan 1994
  • J TROP ECOL
  • 309
  • L J T White
L. J. T. White, J. Trop. Ecol. 10, 309 (1994).
  • Jan 2000
  • CONSERV BIOL
  • 1623
  • J R Malcolm
  • J C Ray
J. R. Malcolm, J. C. Ray, Conserv. Biol. 14, 1623 (2000).
  • Jan 2003
  • 249
  • J S Hall
  • D J Harris
  • V Medjibe
  • P M S Ashton
J. S. Hall, D. J. Harris, V. Medjibe, P. M. S. Ashton, For. Ecol. Manage. 183, 249 (2003).
  • Jan 1999
  • 595
  • J G Robinson
  • K H Redford
  • E L Bennett
J. G. Robinson, K. H. Redford, E. L. Bennett, Science 284, 595 (1999).
Road network upgrading and overland trade expansion in sub-Saharan Africa” (Policy Research Working Paper WPS 4097 World Bank Washington DC
  • P Buys U. Deichmann
  • D Wheeler