98 TRANSPOR TATION RESEARCH RECORD 1291

Secondary and Feeder Road

Development Program
NELSON KuDENGA AND ANDERS BuHRMAN

Research aspects of the Secondary and Feeder Road Develop- • Initiate and introduce alternative design , construction,
ment Program in Zimbabwe are described. The program includes and maintenance concepts for low-volume roads in Zim-
construction, upgrading , and maintenance of approximately 230 babwe;
km of low-volume roads. The main research goals are (a) to
•Evaluate and analyze these concepts from technical,
determine when it may be feasible to seal a gravel road, (b) to
assess the feasibility of an improved gravel wearing course, (c) financial, and economic points of view and arrive at recom-
to evaluate whether unsoaked CBRs could be used as design mendations regarding designs, construction methods , and
CBRs, and (d) to provide recommendations on new design, con- maintenance strategies suitable for similar roads in Zim-
struction , and maintenance methods for low-volume roads in babwe;
Zimbabwe. • Improve or upgrade some existing roads and tracks to
higher standards; and
The Department of State Roads (DSR) within the Ministry •Impart technical knowledge to local Zimbabweans on the
of Transport (MoT) is the road authority in Zimbabwe. DSR design , construction, maintenance , and research methods for
delegates some of its power to rural district councils , town low-volume roads .
councils, and municipalities. The total road network in the
country is over 100 000 km. DSR is in charge of construction It is hoped that the research will answer the following two
and maintenance on all main roads and some secondary and questions:
feeder roads. These roads have been constructed to a high
technical standard, and the road network is considered on e 1. Is it feasible to incre ase the quality of the current gravel
of the best in Africa. wearing course (GWC)?
Soon after gaining independence in 1980, the government 2. At what traffic volumes would it be financially or eco-
of Zimbabwe decided to open up some previously undevel- nomically feasible to upgrade existing earth and gravel roads
oped communal areas . The amount of traffic in these areas using the project concept?
is low (usually less than 100 vpd), but it is increasing as agri-
cultural output increases. Roads in these areas could not all
be built using the traditional high standard for blacktop roads. PROJECT ROADS
Therefore, the government decided to look for alternative
economical ways of building and maintaining low-volume roads Five project roads located in five provinces servicing com-
in communal rural areas using methods and materials suitable munal areas were chosen for the program. These roads are
to the Zimbabwean environment. in different climatic, topographical, and geological environ-
In 1986, an agreement was reached between the govern- ments. New road construction is envisaged for some of the
ments of Zimbabwe and Sweden to implement the Secondary roads. Others will be upgraded to a better standard. Descrip-
and Feeder Road Development Program (SFRDP). This proj- tions of the roads are presented in Tables 1 and 2.
ect was proposed by the Swedish National Road Consulting Originally it was proposed that the project roads include
AB (SweRoad). The study program was presented in 1987 , one or two projects constructed by labor-intensive methods
and the design reports for the roads to be constructed were to allow comparison of total financial costs , economic effects ,
finalized during 1988 and 1989. The construction and upgrad- ·technical standards , and benefits . This part of the project has
ing program started in 1989, two roads will be completed in been delayed but may be included in the final evaluation.
1990, and the remainder will be completed during 1991. One The design, construction , and maintenance of these roads
of the project roads was constructed in 1984 as a demonstra- differ to some extent from the standard approach in Zim-
tion project and is included for maintenance monitoring. babwe, and the actual trials will be carried out on a number
of test sections along the roads.

PROGRAM GOALS
PROJECT ORGANIZATION
The four main goals of this program are to
Because the project contains a large research component ,
N. Kudenga, Department of State Roads, Ministry of Road Trans-
MoT requested and received assistance not only from SweRoad
port, Box 8109, Causeway, Harare, Zimbabwe . A. Buhrman, SweRoad, but also from the Swedish Road and Traffic Research Institute
P.O. Box 8476, Causeway, Harare, Zimbabwe. (VTI) and the Transport and Road Research Laboratory
Kudenga and Buhrman 99

TABLE 1 SFRDP PROJECT ROADS-CHARACTERISTICS

Length Traffic
Road (km) Type (vpd) Remarks
St Joseph-Maphisa Gravel (completed) 60-70 Upgrading including 10-km seal
km 42 + 500-89 + 100 46
Nyika-Zaka Surfaced 200-250 New construction
km 0 + 000-20 + 000 20
Karoi-Binga Surfaced 110-250 New construction/improvements
km 0 + 000-38 + 000 38
km 38 + 000-80 + 000 42 Gravel (completed) 10-50 Upgrading/improvements
Headlands-Mayo 11 + 53 Surfaced/gravel 100 8 km surfaced;
(11 km completed) 2 km gravel; maintenance on
rest
Kamativi-Binga 20 Surfaced (completed 40 Monitoring of deterioration;
1984) maintenance

TABLE 2 SFRDP PROJECT ROADS-ENVIRONMENT

Road Rainfall (mm) Soil Type Economic Activity
St Joseph-Maphisa 400-500 Alkali sand with clay pockets Agriculture ; irrigation schemes;
cattle ranching
Nyika-Zaka 600-800 Gneissic sand with clay and sandy clay Agriculture on subsistence level
Karoi-Binga 800-900 Gneissic sand with clay and sandy clay Agriculture; fishing
Headlands-Mayo 1000-1200 Granitic and gneissic sands and sandy clay Agriculture; livestock
Kamativi-Binga 600-800 Expansive shale soils Fishing; tourism; agriculture on
subsistence level

(TRRL) in the United Kingdom. To facilitate cooperation tions in the unsoaked test results achieved through DCP tests.
between the supporting agencies and MoT staff and to ensure Both of these aspects have been covered in the research pro-
overall control by MoT, project monitoring was delegated to gram and the PME design for improvements. The latter includes
a special unit within MoT-the Project Monitoring and Eval- the use of existing GWC as subbase (Headlands-Mayo) or
uation unit (PME). PME is staffed by engineers and research- base (Kamativi-Binga), assuming unsoaked conditions (4,5).
ers from MoT, the Swedish National Road Administration, The surface of the sealed roads is being constructed with a
and VTI, with short-term input from TRRL. graded gravel aggregate known as Otta-seal or YIG (see Fig-
ure 1). This surface is used mainly in Scandinavia but also on
low-volume roads in Botswana and Kenya (6).
The Otta-seal will provide a dense, flexible cover that,
GENERAL APPROACH AND IMPLEMENTATION depending on the bitumen used, could have self-sealing abil-
ities. Previous experiences in Scandinavia and Botswana indi-
Design cate that the best results are achieved by using a double seal,
with the second seal applied 2 to 15 months later.
The project roads were designed using the existing roads as The pavement design for the gravel roads is new to Zimbabwe
much as possible but aiming at design speeds of 80 or 60 km/ and was based on the specifications presented in Table 3.
hr depending on road type and terrain. Deviations were allowed The GWC used on the project roads should fall as close as
where strict adherence to these design criteria could not be possible to the inner envelope shown in Figure 2, have a PI
justified economically or considering traffic safety aspects. of 10 to 15, and preferably have a clay content of at least 5
The survey and design were carried out using modern instru- percent.
ments and computers and with a minimum of the standard The project GWC is thinner than the standard layer (10 cm
MoT referencing. The design program used was GEO- versus 15 cm) but of considerably higher quality than the
SECMA (1,2). normal MoT standard, by which gravel roads can have a var-
The pavement design for the sealed roads was based on iable surface quality.
MoT recommendations (3), substituting Texas triaxial tests
with California bearing ratio (CBR) values as the main design
criteria and reducing the number of pavement layers to sub- Construction
base and base only. There was a possibility that design CBRs
taken at equilibrium or optimum moisture content (OMC) As mentioned, construction is being carried out using MoT
could be used instead of using soaked CBRs, not only at normal construction and improvement units. Assistance is
designated test sections but also at other places where the provided for planning and follow-up as well as for direct super-
conditions were favorable. However, the design consultant vision of the test sections (e.g., conditional approvals of com-
could not take advantage of this opportunity because of a lack paction are not accepted). Detailed referencing of the road
of reliable data on OMCs and because of unexplained varia- is carried out during the construction stage.
100 TRANSPORTATION RESEARCH RECORD 1291

200 100 so 30 16 B 4 3/8" 3/4"

100

80

l::>
z
VI
VI 60
<t
a..
UJ
l::>
<
h-
z
UJ
u
er 40
UJ
a..

20

0
0.074 0.149 0.297 0.59 1.19 2.36 4.76 9.4 19
SIEVE SIZES (mm)

FIGURE 1 Grading envelope for specification of Otta-seal.

TABLE 3 PAVEMENT DESIGN FOR GRAVEL ROADS same for the gravel roads but are lower for the bitumen roads.
Thickness of
Available figures (excluding Nyika-Zaka) are presented in
Thickness of
Sub grade Wearing Course Selected Fill or Total Thickness Table 4.
Condition (mm) Subbase (mm) (mm)
>SG 9 100 0 100 Maintenance
SG 5 100 125 225
SG 3 100 200 300
The project will include evaluation of different maintenance
<SG 3 100 300 400
policies, especially for routine gravel road maintenance. The
NoTE: Pavement standard design = 0.05 M. SO 5 = subgrade with 5 surface width (6 m) was selected on the basis of studies that
percent < CBR < 9 percent. The subgrade condition will be assessed
from soaked CBR tests carried out at a compacted density of 90 percent used graders with mounted windrow eliminators, which in-
modified AASHTO and at the modified AASHTO OMC. The subgrade creased grading productivity by 60 percent in Zimbabwe (8).
will be compacted to a density of 90 percent or greater of modified AASHTO Another grader attachment used successfully on the improve-
maximum dry density. The selected fill or sub base will be compacted to
a density of 93 percent or greater of modified AASHTO maximum dry
ment projects is the ditch plough, which is mounted directly
density. The wearing course will be compacted to a density of 95 percent on the grader blade and has proven to be very cost efficient
or greater of modified AASHTO maximum dry density. The moisture with loose and normal materials. The windrow eliminator and
content used for compaction of the wearing course and the selected fill ditch plough are now manufactured in Zimbabwe.
or subbase will be within 2 percent of the OMC for the modified AASHTO
compaction. This specification is only applicable for the test sections.
RESEARCH PROGRAM

The maximum stone size for the GWC was set at 20 mm Main Goals
(Scandinavian specifications). Adherence to the envelope will
normally require that the GWC be crushed or screened, or To be able to provide well-founded and objective recom-
both. Therefore, mobile crusher/screener suitable for the con- mendations, a sizable research program was added to the
ditions in Zimbabwe was procured for the project. The impor- project. The program was designed to allow comparisons and
tance of controlling the maximum stone size has been sup- use of results from similar research (e.g., in Kenya, India,
ported by other research in the region (7). In addition to and Brazil). The research will be concentrated on the follow-
crushing or screening, the gravel material must be mixed with ing issues and objectives:
clay in some areas to meet the PI requirement.
Compared with existing MoT standard design and construc- • Determination of when it may be feasible to seal a grav-
tion methods, construction costs for the project are about the el road,
Kudenga and Buhrman 101

100

90

80

70
0\
c
VI
VI
60
Cl
a.
C1J
C1I
50
Cl
~

c
C1J
u
40
'-
C1J
0..
30

20 Ill
~
u.;
~

10

0
0.063 0.212 0.425 2.0 6.3 10 14 20 37.5 63
0.6 5.0 2B
Sieve sizes (mm)

FIGURE 2 Grading envelope for gravel wearing course.

TABLE 4 CONSTRUCTION COSTS BY ROAD TYPE • Reduced thickness of wearing course with impioved-
Width Construction Cost quality material,
Road Type (m) (Z$ per km) • Different maintenance strategies,
• Different traffic levels (depending on the siting of the test
Bitumen MoT design 6/7" 315 000
Bitumen , Project Design 1 617 210 000
sections),
Bitumen, Project Design 2 6/6 110 000 •Different climate (depending on the siting of the test
Gravel MoT 016 110 000 sections), and
Gravel, Project Design 016 110 000 •Use of dust palliatives (i.e., surface binders).
•6-m seal on 7-m formation.
NOTE: Costs reflect 1991 prices. For liitumen roads, Project Design 1 =
standard d ign done by SweR011d and Project De ign 2 = design liy PM - Sealed Roads
on improvement , using unsoaked design BR (normally one tcp lower
in the design table) and without devimions from existing road alignme nt. • Different bituminous seals including priming and surface
treatment,
•Different pavement thickness designs,
•Different quality materials for road base,
• Determination of the feasibility of an improved GWC ,
•Different shoulder treatment,
• Evaluation of whether unsoaked CBRs could be used as
• Different traffic levels (depending on the siting of the test
design CBRs for low-volume sealed roads in certain parts of
sections), and
Zimbabwe,
•Different climate (depending on the siting of the test
• Determination of the feasibility of using DCP tests or the
sections).
Clegg hammer to obtain design CBR values as a supplement
to normal methods ,
Table 5 presents details on variables and measurement methods .
• Evaluation of new construction and maintenance meth-
ods, and
• Development of and recommendations for design charts Test Sections
for low-volume roads in Zimbabwe.
The program includes 26 test sections , of which 9 are located
Variables on gravel road sections, 14 on designed bitumen sections, and
3 on improved sealed sections. The gravel road test sections
The variables to be considered in the research program are (1 km each) are further subdivided into three subsections with
as follows: variable maintenance levels. Most of the test sections are
straight and level with uniform subgrades and include control
Gravel Roads sections of standard MoT design . An example of the test
sections selected for one of the project roads, Headlands-
• Improved-quality wearing course, Mayo, is shown in Figure 3.
TABLE 5 MONITORING OF ROAD SECTIONS
Variable Frequency Method/Equipment Applicable Road Type
Traffic composition Twice/year: 7 days, 6 am to 6 Manual (MoT) G;S"
pm
Traffic volume 7 days, 6 am to 6 pm Manual plus counters G;S
(MoT)
Traffic loading 7 days , 6 am to 6 pm Loadometers (MoT) or G;S
weigh bridges
Traffic speed Before and after maintenance Radar pistol and clocks G;S
Pavement strength and Beginning and end of wet Instructions for G;S
condition season and middle of dry measurements separately
season (G); twice a year issued
(S)
In situ CBR Beginning and end of wet DCP G;S
season and middle of dry
season (G); twice a year
(S)
Density Beginning and end of wet Troxler or CPN Strata G;S
season and middle of dry Gauge
season (G); twice a year
(S)
Moisture content Beginning and end of wet Troxler or CPN Strata G;S
season and middle of dry Gauge
season (G); twice a year
(S)
Impact value Beginning and end of wet Clegg Tmp~ct Tester G
season and middle of dry
season ( G); twice a year
(S)
Indicator tests Immediately after Samples for grading, PI, G
construction, after 1 year, standard tests (in
then yearly transition zones between
test sections)
Deflection Once before construction, FWD G;S
then twice a year (end of
dry season and end of wet
season)
Deformation (G) Before and after maintenance Instructions in appendix G
plus monthly
Corrugation Before and after maintenance Assessment, photos, and G
plus monthly sketch
Potholes Before and after maintenance Assessment, photos, and G
plus monthly sketch
Dusting Before and after maintenance Assessment G
plus monthly
Looseness Before and after maintenance Assessment G
plus monthly
Rutting (maximum Before and after maintenance 2-m straightedge over wheel G
rut depth) plus monthly tracks in several cross
sections along test
sections
Deformation (S) Twice/year Instructions in separate s
report
Cracking Twice/year A frame (1 m 2 ) at each
deflection chainage point.
Rutting (maximum Twice/year 2-m straightedge over wheel
rut depth) tracks in several cross
sections along test
sections.
Surface texture Twice/year Preferably the Mini Texture
Meter (TRRL);
otherwise, UCON sand
patch test in each wheel
track and one in
centerline along each test
section
Roughness Before and after maintenance Bump Integrator and Abay G
plus monthly beam
Roughness Twice first year, then Bump Integrator and Abay s
reassessed beam
Gravel loss • Sections with nil Leveling in cross sections G
maintenance: 3 times/ with optical instruments.
year ( 4-month interval) Cross-sections initially have

TABLE 5 (continued on next page)

Kudenga and Buhrman 103

TABLE 5 (continued)
Variable Frequency Method/Equipment Applicable Road Type
• Sections with one heavy c/c 5 m for comparison.
grading: before and A change to c/c 50 m
after treatment over whole length of test
• Sections with two heavy section might be
gradings: before and proposed after some
after each treatment experience.

Permeability in Once/year Not yet specified s

pavement
Groundwater If necessary Monthly recording of GWL G;S
in pipes in open ground
adjacent to road
Rainfall Daily during wet season Rain gauge installed at G;S
some official buildings
(e.g., a school close to
project roads)
G = gravel roads; S = surfaced roads.
0

NOTE: Visual inspections will also be carried out two or three times per year to subjectively assess the road conditions according to standardized inspection
fo rms used in the 1985 and 1989 Zimbabwe Road Condition Surveys.

TEST SECTIONS ON ROAD 239 HEADLANDS - MAYO (HSMO)

l GRAVEL SURFACED KM 9 -11 I
TEST SECTION NO 1 z
CHAINAGE KM 9 +000- 10 +000 10 +000 - 11+000
LENGTH/WIDTH 1000 M/6M 1000M/6M
SURFACING WC 100 MM ACC. TO PROJ . WC 100 MM ACC. TO MOT

SUBBASE/SELECTED FILL 175 MM 175 MM

SUBGRADE SG' 5 SG 5

II GRA~INGS
1 II II
MAINTENANCE/ LENGTH NIL
lOOM
II GRA 01NG
lOOM 300 M
ACC. TO SECTION r
II II

!SEALED SURFACED KM 0 - 9 I

TEST SECTION NO r
CHAINAGE KM 5 +800 - 6 +300
LENGTH/ WIDTH SOOM/6M

SURFACING SINGLE OTIA SEAL

BASE 150MM
SUB BASE 0

SUBGRADE SG9

FIGURE 3 Test sections on project road.

Detailed preinvestigations have been performed on all test research on the gravel roads should be complete and a pre-
sections. The preinvestigations included soil classifications, liminary report on the sealed roads should be made available.
layer thicknesses, in situ and laboratory CBR tests, mois- The final report on the sealed roads cannot be expected
ture contents, PI values, Clegg hammer and DCP values, until at least 5 years after the roads have been constructed,
falling weight deflectometer (FWD) data , and roughness data and the final answers on some questions may only be available
over time. after 8 to 10 years or when a complete overlay is required.

Time Schedule and Resources Intermediate Conclusions

The field measurements will require at least three field assist- From the preinvestigations and subsequent monitoring of
ants supported by two full-time researchers in the PME during the test sections, the following preliminary conclusions have
the first phase, which should end in mid-1992. At that time, been made:
104 TRANSPORTATION RESEARCH RECORD 1291

•The lowest in situ CBR values (down to 600 mm) gen- mented during 1990 and 1991 will be presented at the TRB
erally exceed the laboratory soaked CBR values, indicating 5th International Conference on Low-Volume Roads.
that a pavement design that is based on the latter would be
very conservative. The results tend to support upgrading the ACKNOWLEDGMENTS
subgrade classification by at least one step (9). This project was funded by the Swedish International Devel-
• Current design methodology, which is based nn a limited opment Agency. The authors are grateful to the Zimbabwe
number of soaked CBR values and grouping of sub grade sam- Secretary for Transport for his permission to publish this paper.
ples, is unreliable (9).
• A design methodology that is based on a large number REFERENCES
of DCP values taken in March and assessed statistically would
be more informative for the designer (9). 1. Secondary and Feeder Road Development Programme (SFRDP).
Study Report. Ministry of Transport, Zimbabwe; SweRoad/Sweco,
• A good correlation between DCP and CIV has only been Sweden, 1987.
established in sand. For other materials, it has not been pos- 2. SFRDP: Design Reports, 8 vols. Ministry of Transport, Zim-
sible to establish a general relationship (10). babwe; SWECO, Sweden, May-July 1988.
•The performance of the Kamativi-Binga and Headlands- 3. R. L. Mitchell. Flexible Pavement Design in Zimbabwe. Proc.,
Mayo roads supports the view that low-cost sealed roads are Institute of Civil Engineering, Vol. 72, Part 1, 1982.
4. SFRDP: Headlands-Mayo, Improvement Report. Ministry of
more cost efficient than standard gravel roads in Zimbabwe Transport, Zimbabwe; SWEROAD, Sweden, May 1990.
(4-6). 5. B. Lindberg. An Assessment of the Pavement Structure and Strength
of the Road Kamativi-Binga. University of Technology, Lund,
Sweden, 1991.
6. Performance Update on Low Cost Sealed Roads. Roads Depart-
EVALUATION
ment, Botswana/NRRL, Norway, 1984.
7. P. Paige-Green. The Influence of Geotechnical Properties on the
Evaluation of the alternatives will be based both on perfor- Performance of Gravel Wearing Courses. University of Pretoria,
mance (i.e., assessments of standard and rate of deterioration) Union of South Africa, 1989.
and on financial and economic costs. The economic costs will 8. Work Study on the Use of Windrow Eliminator. Ministry of Trans-
port, Zimbabwe, 1985.
be calculated using both the HDM III and RTIM 2 computer 9. SFRDP: Discussion of CBR and Moisture Content Results. Work-
models. In addition, checks will be made with the NTS (11). ing Paper 2. Ministry of Transport, Zimbabwe; SWEROAD,
Accident costs will probably not be included in the economic Sweden, July 1990.
calculations because their influence will likely be marginal. 10. J. Dahlberg. Pavement Strength and Pavement Conditions on a
Gravel Road in South Western Zimbabwe. Royal Institute of
The more detailed performance evaluations and individual Technology, Stockholm, 1990.
relationships found may possibly be reported continuously, 11. Zimbabwe National Transport Study. Ministry of Transport, Zim-
at least on the gravel road trials. Findings and results docu- babwe; SWECO, Sweden, 1985.