American Journal of Plant Sciences, 2010, 1, 38-46
doi:10.4236/ajps.2010.11006 Published Online September 2010 (
Copyright © 2010 SciRes. AJPS
Nutrient Flows in Perennial Crop-Based Farming
Systems in the Humid Forests of Cameroon
Eugene Ejolle Ehabe1*, Nomo Lucien Bidzanga2, Charles-Magloire Mba1, Jetro Nkengafac Njukeng1,
Inacio de Barros3, Frank Enjalric4
1Institute of Agricultural Research for Development (IRAD), Ekona Regional Research Centre, PMB 25 Buea, Cameroon; 2Institute
of Agricultural Research for Development (IRAD), Nkolbisson, Yaounde, Cameroon; 3INRA Antilles-Guyana, URAPC, Guadeloupe
(FWI); 4Unité Mixte de Recherche Système, CIRAD Cultures Pérennes, Montpellier, France .
Email: *
Received August 4th, 2010; revised August 24th, 2010; accepted September 11th, 2010
A study was conducted in some perennial crop-based farms in the humid forests of South West Cameroon, to better un-
derstand their soil fertility patterns and management and identify factors that contribute most to nutrient depletion
trends in such multi-storey farmholdings. The main pe rennial crops were the para rubber tree (Hev ea brasiliensis), co-
coa (Theobroma cacao) and oil palm (Eleais guineensis) whose ages ranged from immature to very old (senescence).
Data were collected over a two-year period (2007 and 2008) and modelled using the farm NUTrient MONitoring
(NUTMON) tool. Results showed that the farming systems played key roles in the overall exploitation strategies. Plan-
tain and cassava (annual crops) and cocoa (perennial) were the most associated crops, accounting for more than half
of intercropped frequ encies. Whereas nutrient flows within an d between farm units were confirmed , the net loss in soil
N, P and K nutrient balances differed with the main perennial crop - being highest for the oil palm and lowest for the
rubber tree. The average nutrient balance of each farm was markedly negative for N but positive for P and K. Partial
nutrient balance (ignoring biophysical flows like N-fixation, leaching and erosion) was positive for the three nutrients
indicating therefore that the biophysical flows accounted more for N depletion. Nitrogen loss was mainly due to, leach-
ing (~70%), volatilization (~20%), and exported crops and their residues.
Keywords: Soil Fertility, Nutrient Monitoring, Depletion, Nutmon
1. Introduction
Crop production practices within peasant farm holdings
in sub-Saharan Africa often progress at the expense of su-
stainable land use as farmers are primarily concerned ab-
out crop and animal production, for the forthcoming sea-
son. Hence, long-term processes that adversely affect su-
stainability, such as decrease and eventual depletion of
soil nutrient stocks, are less visible and receive a lower
priority. Most often, these soils are progressively being
mined of their nutrients [1] whereas there is need to se-
cure present productivity and ensure the sustainability of
these farming systems. These farmers have to manage the
fertility of their soils by manipulating the flows of nutri-
ents into, out of and within the farm. Indeed, decisions
concerning soil productivity are determined by house-
hold objectives (food security, profit or cash maximiza-
tion, risk aversion, long term security of livelihood , etc.),
available resources (labour, cash, implements, natural re-
sources, etc.) and the socio-economic environment.
Soil fertility in sub-Saharan Africa is today regarded
as a major survival issue, especially in peasant perennial
crop-based systems in the humid forests of Cameroon.
This project was initially undertaken to ensure a better
understanding and improvement of soil fertility in these
cropping systems, to determine the possible role of integ-
rated nutrient management in reversing present trends of
nutrient depletion at each farm-level, and to integrate qu-
antitative and qualitative assessment approaches in man-
aging soil fertility.
The complexity of peasant farms does not always all-
ow for quantification of all flows and stocks. Nutrient
monitoring (NUTMON) is a multi-scale modelling app-
roach that was conceived to assess stocks and flows of
certain nutrients in a well defined geographical unit foll-
owing a strict monitoring of all farm inputs (mineral fer-
tilizers, manures, atmospheric deposition and sedimenta-
tion) and outputs (h arvested crop produces, residues, lea-
ching, denitrification, erosion losses). The NUTMON-
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
Toolbox was developed with three modules to quantify
nutrient flows, nutrient stocks and economic performance
indicators for farms. It provides guidelines for data gath-
ering in the field, data entry, calculation of flows and
balances and data interpretation. Application of the tools
from the toolbox results in a thor ough analysis providing
insight into nutrient flows between the various activities
within the farm as well as between the farm and its sur-
roundings [2].
The specific objectives of this study were therefore to
evaluate variations of soil nutrient stocks at farm level,
measure and estimate the major nutrient flows at farm
level as a means to describe current farm nutrient manag-
ement, to identify the k ey factors influencing land mana-
gement in perennial crop-based farms in the humid fore-
sts of South West Cameroon in a bid to discover some of
the underlying causes of so il fertility depletion. This was
particularly supposed to be done using the Nutrient Mo-
nitoring Programme (NUTMON) or software, a tool that
was developed to assess nutrient balances (stocks and fl-
ows of some macro-nutrients - N, P and K), biomass fl-
ow and economic performance at farm level [2].
NUTMON has been suggested as a decision-support
model to monitor the effects of chang ing land us e and su-
ggest interventions that improve the nutrient balance [4].
As input and output determinants can not all be qu antified
equally well, the model recognizes primary data, estima-
tes and assumptions. The NUTMON determinants are
mostly scale-neutral and can therefore be used to monitor
nutrient balances at farm, regional, national and supra-
national levels.
NUTMON is fed by a number of basic data, and by
nutrient input and output data [2]. Basic data include the
surface area of the arable land, and the spatial patterns of
land use systems (LUS). Nutrient input and output data
are reflections of different processes, each of which has a
certain value (the nutrient balance) which is specific for a
given LUS at a given time. A second monitoring exercise
at a later stage may yield different results, which may be
due to changes in the LUS, or changes in the individual
nutrient input and output values. As the changes have eit-
her aggravated or ameliorated the nutrient balance, NU-
TMON can support decision-making in the interest of
sustainable forms of agriculture [3,5].
2. Methodology
2.1. The Study Area
Field observations were carried out on 19 typical peren-
nial crop-based agroforestry farms, distributed around Ku-
mba and extending to the Bombe-Malende zones (4°25’-
4°80’N and 9°25’- 9°35’E) in the South West Region of
Cameroon (Figure 1). This region falls within the rain-
forest area (mean rainfall of 2,852 mm/yr), with a mark-
ed rainy season (March to October), and high mean ann-
ual temperatures (~23˚C) evenly distributed throughout
the year [6], Soils are ferrallitic with patches of fertile vol-
canic areas, and altitudes varying from 25 to 400 m to-
ward the North. The agro-forestry exploitations existing
in the area are typically characterized as home gardens
[7], that is, permanently occupied (no fallow) small areas,
usually adjacent to the farmers’ houses, integrating main
perennials (oil palm, rubber trees and cocoa), food crops
(yams, cassava, maize, banana, plantain, cocoyam, etc.),
and native trees (as well as ornamentals and medicinal
plants not considered in the surveys).
The farm holdings had been judged homogeneous in
terms of their farm and non-farm economic activities as
well as the presence of some perennials as main crop in
association with other crops. These farm types ranged
from small-scale holdings with mostly food crops, indus-
trial plantations for export crops and intercropped farms
having various combinations of perennial and food crops
[8,9]. The main crops, their development stages and basic
characteristics of the terrain of the studied farms are pre-
sented in Table 1. The parent material of all the so ils was
of a volcanic/basement complex [1,10].
Prior to the survey, each farm was characterised for its
location, altitude, rainfall and surface area, relief, soil mo-
rphology (depth, texture or clay content, colour, parent
material) and composition (moisture status, rooting status,
2.2. Soil Analyses
For each plot, three to five representative samples were
collected, following the procedure described latter on,
and bulked. A soil observ ation was made at 100-m inter-
vals using a soil auger to examine the soil down to a de-
pth of 50 cm or more, when feasible. During such obser-
vations, several soil/land characteristics were noted, no-
tably the slope, horizon differentiation, soil depth, texture,
colour, coarse fragments, drainage conditions, etc. These
were later on sorted into groups of similar soils and ter-
rain. Similar soils were bulked, labelled and packaged for
eventual laboratory analysis.
In the laboratory, the soils were dried under ambient
conditions for about 2 weeks, crushed and siev ed through
a 2-mm mesh. The throughs were subjected to several
standard soil analytical procedures, notably for their:
- texture (particle size) using the pipette method to ob-
tain the relative proportions of sand, silt and clay;
- pH (water and KCl), measured potentiometrically in
a soil suspension (1:2.5) after 24 h;
- Total Nitrogen (N) by the Kjeldahl semi-micro meth-
od [11];
- Available Phosphorus (P) by the Bray II me t hod [ 12] ;
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
Figure 1. Map of Cameroon showing the study site.
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
Table 1. Farmholdings covered during the study.
Plot Village Main crop Development of main crop Altitude (m) Slope (%) Clay (%) FSU
18 Bombe Oil palm In production 208 1 35 A
21 Bombe Cocoa Immature 45 3 30 A
22 Malende Cocoa Immature 67 5 35 A
23 Bombe Cocoa Immature 89 3 40 C
24 Malende Cocoa Immature 38 3 30 A
25 Bombe Cocoa In production 45 0 40 C
26 Bombe Cocoa In production 28 0 40 C
27 Bombe Cocoa In production 56 20 40 D
28 Malende Oil palm Immature 87 13 30 B
29 Malende Oil palm Immature 56 3 30 A
30 Bombe Oil palm Immature 34 4 30 A
31 Bombe Oil palm In production 25 5 30 A
32 Bombe Oil palm In production 56 3 25 A
33 Bombe Oil palm In production 85 3 30 A
34 Bombe Rubber Immature 82 3 30 A
35 Bombe Rubber Immature 67 1 30 A
36 Bombe Rubber Immature 65 2 30 A
37 Bombe Rubber In production 69 1 30 A
38 Bombe Rubber In production 70 25 40 D
- Exchangeable bases calcium (Ca), magnesium (Mg),
potassium (K) and sodium (Na) by the neutral ammo-
nium acetate leaching method [13,14];
- Aluminium and hydrogen ions (Al+ H) by leaching
with 1M KCl [13,14];
- The soil bulk dens ity was determined from soil cores
using the Cylinder method [15,16 ]. Soil samples were dr-
ied at 105˚C to constant weight, and th e bulk density was
recorded as the ratio of the mass of the dried so il (in g) to
the internal volume of the core cutter (in cm3).
2.3. Data Collection and Analysis Using
The NUTMON ® tool, applied at both the farm and ac-
tivity levels, included a structured questionnaire, a data-
base, and two static models - NUTCAL for calculation of
nutrient flows and the ECCAL for calculation of econ-
omic parameters [17]. A user interface facilitated data
entry, data manipulation and data extraction to produce
input for both models.
An initial inventory of household composition, farm
and field layout, agricultural activ ities an d nutrient sto cks
was completed for each farm. This inventory was follo-
wed by a monthly monitoring of on-f arm agricultural ac-
tivities that affect nutrient flows. The activities of the fa-
rm family, their cash income and the allocation of labour
were also recorded.
Based on the algorithm of the NUTMON ® software,
the obtained data were modelled to obtain nutrient bala-
nces (inputs less outputs) for all the farms. Nutrient outp-
ut included harvested crop products and residues, nutria-
nts leached out below the root zone, gaseous losses from
the top soil, erosion and human excreta lost because it
ends up in deep pit latrin es far below the root zone. Nut-
rient inputs into the system were mineral fertilisers, org-
anic inputs (manure, imported crop residues and feeds),
air-borne deposition, biological N-fixation by plants, sed-
imentation and nutrients extracted from the sub-soil by
deep-rooting crops and trees.
3. Results and Discussion
3.1. Characterisation of Farm Sections
Initial assessment as shown in Figure 2 (not derived us-
ing the NUTMON ®), showed that plantain and cassava
were the most intercropped annual crops while cocoa,
and to a slightly lesser extent, o il palm were the most in-
tercropped perennial crops in the study area. In effect, the
two respective crops in each categ ory accounted for more
that 50% of observed intercropped frequencies.
Interviews confirmed that the farmers were aware of
the progressive decline in land productivity over the yea-
rs, a phenomenon attributed to continuous cropping on
the same fields, soil erosion and most especially the dim-
inishing supplies of manure. However, they were not aw-
are of the fate of macro nutrients in these processes and/
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
or how this affected their soil productivity.
3.2. Analysis of Nutrient Flows in Farm Units
The overall N-P-K nutrient flows showed that these flo-
ws varied with the n utrien ts un der consid eratio n an d with
the distribution factors considered (Figure 3).
N, P and K in-flows seemed to be evenly distributed
between imported organic fertiliser (IN1a), wet and dry
atmospheric deposition (IN3), N-fixation (IN4) and to a
lesser extent the minerals in harvested produce. This, ho-
wever, was not the case for out-flows as the proportions
varied considerably with the nutrients considered. First,
the greatest nutrient loss was that of nitrogen, and this
was mainly attributed to leaching (~70%), volatilization
(~20%), and to exported crops and crop residues. Very
little of P and K were lost from the farms. Instead, these
Cas sava
Frequency (%)
Oil palm
Cas hew
Ma ngo
Frequency (%)
Figure 2. Frequency of occurrence of the main (a) annual crops and (b) perennial crops associated in the perennial crop
based smallholdings.
Mineral fertiliser, IN1
Imported organic
fertilise r, IN2a
Imported manure,
At m ospheric
depos it ion, IN3
N-fixation, IN4
Harvests, OUT1
Exported crop
residues, OUT2a
Manure excr etion
outside f arm, OUT 2b
Leaching, O UT 3
Gaseous loss, O U T4
E rosi o n, O UT5
Qua ntity o f nutrients tra nsfe r re d,
Full balancePartial balance
Nutrients transferred ( kg/ha)
Figure 3. Overall N-P-K nutrient flows for the farms studied, as well as the specific full and partial balance of N-P-K nutrient
transfers (insert).
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
nutrients were mostly obtained from imported organic fe-
rtilisers and atmospheric deposition. These figures were
confirmed by the rather different overall balances, nega-
tive for nitrogen and positive for phosphorus and potass-
ium (Figure 3(a)). On ignoring all losses due to leaching
(partial nutrient balance), the overall figures became pos-
itive for the three elements, especially for N (Figure 3,
insert). This demonstrated that farmers may have inadv-
ertently compensated for depletion provoked by harvest-
ing of farm produce by returning global biomass at the
plot level.
3.3. Nutrient Flows between Farm Units
Full and partial balances of the nutrients showed larg e di-
fferences between farm units (Figure 4). Whereas all fa-
rm units had negative full balances for nitrogen (Figure
4(a)), only about one-quarter of them had a negative par-
tial balance for that nutrient (Figure 4(b)). This situation
was slightly different for K transferred, but much differ-
ent for P. Indeed, almost all the farm units had positive
full and partial balances for P wh ile a small proportion of
them continued to have negative values for K for both
the full and partial balances. These results indicate
clearly that leaching might be the most important factor
contributing to loss of soil nutrien ts, especially nitrogen.
When all the farms were considered as a farm section
unit (FSU), the contribution of various modelled factors
to the overall nutrient status (Figure 5) as well as to the
full and partial nutrient balances (Figure 5, insert) chan-
ged considerably. From these, it could be shown that at-
mospheric deposition was responsible for most of the N-
nutrient transferred into the FSU, followed by N fixation
and imported organic fertiliser. Once more, leaching and
gaseous loss accounted for most of the N loss while har-
vested produce and the exported crop residues accounted
for P and K loss.
Net nutrient losses in soil N, P and K were much more
pronounced in cocoa and oil palm based fields than the
Hevea ones. This will certainly be related to the multi-
plicity of products harvested from the oil palm and cocoa
fields compared to the rubber fields which give essen-
tially C, captured though photosynthesis.
21 22 23 24 25 26 27 28 2930 31 32 33 34 35 36 37 38
Full balance of nutrients transferred, kg/h
- N -- P -- K -
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
Partial balance of nutrients transferred, kg/h
- N -- P -- K -
Figure 4. Specific full and partial balance of overall N-P-K nutrient transfers for the individual farm units.
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
Mineral fertiliser,
Imported organic
fertiliser, IN2a
Atm osph eric
deposition, IN3
N-fixation, IN4
Harvests, OUT1
Exported crop
residues, OUT2a
Leaching, OUT3
Gaseous loss,
Erosion, OUT5
N and P transferred, g/ha
K transferred, g/ha
Nutrient balance, g/ha
Full Pa rtial
Figure 5. Contribution of various fac tor s to the N-P-K nutrient balance in the farm section unit (FSU) and their contributions
to the total nutrient balance (insert).
Table 2. Flows per unit outside the farm for the entire farm section unit.
(a) Rubber farm
Movement Material Fresh, kg/haDry, kg/ha N, kg/haP, kg/haK, kg/ha
OUT Rubber coagula 737 295 0.8 0.3 0.8
OUT Pineapple fruit 4 1 0.0 0.0 0.0
IN Formic acid 4 4 0.0 0.0 0.0
IN Ethrel 1 1 0.0 0.0 0.0
(b) Oil palm farm
in/out material Fresh, kg/ha Dry, kg/ha N, kg/haP, kg/ha K, kg/ha
OUT Cassava - tuber 467 177 0.7 0.1 0.7
OUT Oil palm nu t s 459 459 10.7 15.6 20.9
OUT Palm oil 3560 3560 12.1 160.2 12.7
OUT Pumpkin – seeds 80 72 0.6 0.1 0.2
IN Calash 1 1 0.0 0.0 0.0
IN 20/10/10 5 5 1.0 0.2 0.4
IN Maize - grain 4 3 0.1 0.0 0.0
IN Okra - grain 0 0 0.0 0.0 0.0
IN Plantain - suck e r 70 70 0.3 0.3 0.1
IN Yam - tuber 23 6 0.0 0.0 0.0
IN Pumpkin - gr ain 11 10 0.1 0.0 0.0
IN Insecticides 1 1 0.0 0.0 0.0
IN Plantain - bunch 23 4 0.0 0.0 0.0
IN Maize – cob 30 26 0.4 0.1 0.1
IN Taro/cocoyam tuber 15 15 0.1 0.1 0.2
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
(c) Cocoa farm
in/out Material Fresh, kg/haDry, kg/haN, kg/ha P, kg/ha K, kg/ha
OUT Cocoa - grain 122 122 4.0 5.6 0.4
OUT Pineapple - fruit 3 1 0.0 0.0 0.0
OUT Plantain - bunch 3681 626 0.8 0.2 2.8
OUT Rice - grain 60 54 0.6 0.2 0.2
OUT Cassava - processed 70 27 0 .1 0.0 0.1
OUT Groundnuts 220 209 7.8 1.3 1.7
IN Potato – tuber 10 2 0.0 0.0 0.0
IN Cocoa - seedlings 450 450 9.0 2.1 1.5
IN Rice - grain 68 61 0.7 0.2 0.2
IN Cassava - tuber 78 29 0 .1 0.0 0.1
IN Maize – cob 338 294 4.9 1.2 1.4
IN Plantain - sucker 395 395 1.4 1.8 0.7
IN Cocoa – seeds 5 60 560 18.5 25.8 2.0
IN Taro/cocoyam tuber 50 50 0.2 0.3 0.6
Economic analyses, which formed a related aspect of
this study but presented elsewh ere [18], show ed that cash
crops, as compared to staple crops, could be more profit-
able and often associated with less negative nutrien t bala-
nces than food crop because more inputs were invested in
them (Table 2). At farm level, a higher nutrient loss was
calculated for the more heavily intercropped fields, nota-
bly oil palm and cocoa. Indeed, these fields suffered the
greatest nutrient depletion due to the multiple forms in
which their products were harvested. As shown in Table
2, the Hevea fields mostly lost their nutrients following
tapping and draining of their latex whereas the oil palm
lost its nutrients through its harvested bunches (seeds,
kernels and leaves ).
As our results would indicate, nutrient balance models
are valuable when using data obtained from a site or set
of sites to extrapolate to larger areas [4]. However, there
are limitations inherent to su ch aggregations, notably wi-
th respect to the representativeness of sample site(s) and
the applicability of data collected at one scale (geograp-
hic or temporal) in representing processes, which may oc-
cur at broader scales [19]. NUTMON is a multi-scale ap-
proach that assesses stocks and flows of some macro-nu-
trients (N, P and K only) in a well defined geographical
unit. It is often considered as pragmatic to carry out nu tr-
ient audits after calculating nutrien t balances and evalua-
ting trends in nutrient mining/enrichment [20]. The NU-
TMON model has therefore been widely adopted though
the several major simplifications (coefficients for crop
residue removal are the same for all crops) and its inabil-
ity to measure effects of interactions between the nutri-
ents seem to make the results less reliable [21,22]. Furth-
ermore, the model ignores data on carbon sequestration,
a factor that would much improve its usefulness and app-
4. Conclusions
Overall N-P-K flows varied with the nutrients an d the d is-
tribution factors considered. The in-flow of these nutria-
nts was evenly distributed between dry atmospheric
deposition, N-fixa tion and the harvested produce. Nitrog-
en was the nutrient lost in greatest amounts, due essenti-
ally to leaching, gaseous loss and harvesting of crops and
their residues. Less important amounts of P and K were
lost from the farms as substantial amounts were obtained
from imported organic fertilisers and atmospheric depos-
ition. Large differences were observed between farm
units with respect to their full and partial balances of the
nutrients. Despite the sustained productivity within pere-
nnial crop-based cropping systems, all farm units had ne-
gative balances for N and almost none for P and K. Net
nutrient losses in soil N, P and K were much more pron-
ounced in cocoa and oil palm based fields than the Hevea
ones, probably because of the multiplicity of products
harvested from former fields. The more heavily interc-
ropped fields were subjected to greater nutrient depletion.
These results could nonetheless be treated with some ca-
ution and the Nutmon software revised to eli minate its se-
veral major simplifications, its inability to measure effe-
cts of interactions between the nutrients, and its non-uti-
lisation of carbon sequestration data, all of which seem to
make the results less reliable.
5. Acknowledgements
This work was undertaken in the framework of the ATP
CARESYS Project and funded by the Centre Internatio-
nal de Coopération en Recherche Agronomique pour le
Développement (CIRAD). The Logistic and other support
of the Regional Directorate of CIRAD for Central Africa
are highly appreciated.
Nutrient Flows in Perennial Crop-Based Farming Systems in the Humid Forests of Cameroon
Copyright © 2010 SciRes. AJPS
[1] J. Kanmegne, “Slash and Burn Agriculture in the Humid
Forest Zone of Southern Cameroon: Soil Quality Dyn-
amics, Improved Fallow Management and Farmers’ Per-
ceptions,” PhD Thesis, Wageningen University and Re-
search Centre, 2004.
[2] H. V. D. Bosch, A. de Jager and J. Vlaming, “Monitoring
Nutrient Flows and Economic Performance in African
Farming Systems (NUTMON) II. Tool Development,”
Agriculture, Ecosystems and Environment, Vol. 71, No. 1,
December 1998, pp. 49-62.
[3] A. de Jager, S. M. Nandwa and P. F. Okoth, “Monitoring
Nutrient Flows and Economic Performance in African
Farming Systems (NUTMON) I. Concepts and Method-
ologies,” Agriculture, Ecosystems and Environment, Vol.
71, No. 1, December 1998, pp. 39-50.
[4] E. M. A. Smaling and L. O. Fresco, “A Decision Support
Model for Monitoring Nutrient Balances under Agricul-
tural Land Use (NUTMON),” Geoderma, Vol. 60, No.
1-4, December 1993, pp. 235-256.
[5] U. Surendran, V. Murugappan, A. Bhaskaran and R. Jaga-
deeswaran, “Nutrient Budgeting Using NUTMON-Tool-
Box in an Irrigated Farm of Semi-Arid Tropical Region in
India: A Micro-and Meso-Level Modelling Study,” World
Journal of Agricultural Sciences, Vol. 1, No. 1, January
2005, pp. 89-97.
[6] E. E. Ehabe, M. T. Besong and S. W. Almy, “Late and
Infrequent Weeding by Peasant Farmers in the Humid
Forest of Cameroon,” Tropical Science, Vol. 41, No. 3,
September 2001, pp. 137-141.
[7] F. L. Sinclair, “A Genera l Classification of Agro-Forestry
Practice,” Agroforestry Systems, Vol. 46, No. 2, Septem-
ber 1999, pp. 161-180.
[8] S. J. Epale, “Plantations and Development in Western Ca-
meroon. 1885-1975. A Study in Agrarian Capitalism,”
Vantage Press Inc., New York, 1985.
[9] C. Plaza “Situation de l'hévéaculture Villageoise dans une
Filière en Évolution: Cas de la Région d'Ekona, Province
du Sud-Ouest Cameroun,” M.Sc. Thesis, Université Mon-
tpellier I, France, 2003.
[10] M. T. Besong, S. W. Almy and B. Bakia, “Land Produc-
tivity of Small Farm Holders in South West Province of
Cameroon,” International Journal of Tropical Agricul-
ture, Vol. 11, No. 1, March 1993, pp. 81-87.
[11] J. Jones and J. Benton, “A Kjeldahl Method of Nitrogen
Determination,” Micro-Macro Publishing Inc., Athens, Ge o r -
gia, USA, 1991.
[12] R. H. Bray and L. T. Kurt, “Determination of Total Org-
anic Available Forms of Phosphorus in Soils,” Soil Sci-
ence, Vol. 59, No. 1, January 1945, pp. 39-45.
[13] C. A. Black, “Methods of Soil Analysis,” American Soci-
ety of Agronomy, No. 9, Part 2, Madison, Wiscosin. 1965.
[14] M. L. Jackson, “Soil Chemical Analysis,” Prentice Hall,
New York, 1962.
[15] A. L. Tisdall, “Comparison of Methods of Determining
Apparent Density of Soils,” Australian Journal of Agri-
cultural Research, Vol. 2, No. 4, December 1951, pp.
349- 354.
[16] B. R. Brasher, D. P. Franzmeier, V. Valassis and S. E. Da-
vidson, “Use of Saran Resin to Coat Natural Soil Clods
for Bulk Density and Moisture Retention Measurements,”
Soil Science, Vol. 101, No. 2, February 1966, pp. 108-
[17] J. Vlaming, A. de Jager, R. van den Bosch, G. Mei-
jerink, C. van Beek , S. van Wijk and H. v an Ke ul en ,
NUTMON Toolbox-Data Background, Data Entry,
and Data Processing Modules, v3.5-1. "Atambua".
Jointly published by Envista Consultancy, Agricul-
tural Economics Research Institute (The Nether-
lands), Alterra Consultancy and Plant Research In-
ternational (The Netherlands), 2006. http://www
[18] G. S. Rodrigues, I. de Barros, E. E. Ehabe, P. Sama-Lang
and F. Enjalric, “Integrated Indicators for Performance
Assessment of Traditional Agroforestry Systems in South
West Cameroon,” Agroforestry Systems, Vol. 77, No. 1,
September 2009, pp. 9-22.
[19] P. Drechsel, M. Giordano and L. Gyiele, “Valuing Nutri-
ents in Soil and Water: Concepts and Techniques with
Examples from IWMI Studies in the Developing World,”
Research Report 82, Colombo, International Water Man-
agement Institute, Sri Lanka, 2004.
[20] U. Surendran and V. Murugappan, “Pragmatic Approach-
es to Manage Soil Fertility in Sustainable Agriculture,”
Journal of Agronomy, Vol. 9, No. 2, January 2010, pp.
[21] N. R. Rabindra and R. V. Misra, “Soil Nitrogen Balance
Assessment and its Application for Sustainable Agricul-
ture and Environment,” Science in China Series C - Life
Sciences, Vol. 48, No. 6, January 2005, pp. 843-855.
[22] U. Surendran and V. Murugappan, “Nutrient Budgeting
in Tropical Agro Ecosystem-Modeling District Scale Soil
Nutrient Balance in Western Zone of Tamil Nadu Using
Nutmon-Toolbox,” International Journal of Soil Science,
Vol. 2, No. 3, September 2007, pp. 159-170.