Soil erosion from shifting cultivation and other smallholder

land use in Sarawak, Malaysia

Andreas de Neergaard a,*, Jakob Magid a, Ole Mertz b

a Department of Agricultural Sciences, Faculty of Life Sciences, University of Copenhagen, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmarkb Department of Geography and Geology, Faculty of Science, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark

Received 23 August 2007; received in revised form 17 December 2007; accepted 18 December 2007

Available online 13 February 2008

Abstract

The sustainability of shifting cultivation systems and their impact on soil quality continues to be debated, and although a growing body of

literature shows a limited impact on, e.g. soil carbon stocks, shifting cultivation still has a reputation as detrimental to the environment. We

wished to compare soil erosion from three land use types in a shifting cultivation system, namely upland rice, pepper gardens and native forest.

We used two sample sites within the humid tropical lowland zone in Sarawak, Malaysia. Both areas had steep slopes between 258 and 508, and

were characterised by a mosaic land use of native forest, secondary re-growth, upland rice fields and pepper gardens. Soil samples were

collected to 90 cm depth from all three land use types, and analysed for various chemical parameters, including texture, total organic matter

and 137Cs content. 137Cs is a radioactive isotope derived from nuclear fallout, and was used to estimate the retention of topsoil in the profiles.

Soil chemical parameters in upland rice fields, such as extractable cations, pH and conductivity, indicated limited soil transportation

downslope, and depletion of cations from upslope samples are most likely caused by leaching and losses via ashes after clearing and burning.

The position on slope had no significant effect on soil texture, carbon or P content, indicating very limited physical movement of soil

downslope. A soil carbon inventory to 90 cm depth on the three land uses only showed a higher carbon concentration in the top 5 cm of forest

and upland rice plots. When corrected for soil density, there was no effect of land use on the carbon inventory. Moreover, the carbon content in

the top 30 cm contributed <50% of the total carbon inventory, hence even significant effects of land use on carbon content in the upper soil

layers, are unlikely to change the carbon inventory dramatically. 137Cs content in the soil profile indicated largest retention of original topsoil

in the native forest plots, and a loss of 18 and 35% of topsoil from upland rice and pepper gardens, respectively, over the past 40 years. When

comparing to 30 cm depth, soil loss was 30% from both upland rice and pepper fields. Low 137Cs activity in deeper soil layers rendered a total

profile inventory impossible. It is concluded that shifting cultivation of upland rice in the current system is not leading to degradation of soil

chemical and physical quality. The soil carbon inventory is not affected by land use in this analysis, due to the contribution from the deeper soil

layers.

# 2008 Elsevier B.V. All rights reserved.

Keywords: 137Cs; Erosion; Upland rice; Black pepper; Soil carbon; Slash-and-burn; Swidden farming

1. Introduction

Soil erosion and loss of soil carbon is repeatedly being

mentioned as a global threat to environment and food supply

(Pimentel et al., 1995; Pimentel, 2006), but the available

knowledge on erosion in smallholder farming systems is

relatively limited. Shifting cultivation has for decades been

* Corresponding author. Tel.: +45 35333499; fax: +45 35333460.

E-mail address: adn@life.ku.dk (A. de Neergaard).

0167-8809/$ – see front matter # 2008 Elsevier B.V. All rights reserved.

doi:10.1016/j.agee.2007.12.013

accused of causing soil erosion, and already more than 30

years ago Lal (1974) noted that the topic was frequently

discussed in the literature although little empirical evidence

was available. The problem does, like other conventional

understandings of shifting cultivation (Mertz, 2002), prevail,

and many studies still refer to soil erosion and fertility loss as

one of the main environmental and productivity problems in

shifting cultivation (Brady, 1996; Harwood, 1996; Devendra

and Thomas, 2002; Borggaard et al., 2003; Rasul et al.,

2004)—a view often criticised for being far too simplistic

(Kleinman et al., 1995; de Jong, 1997; Schmidt-Vogt, 1998;

Fox, 2000). Some empirical studies are now available, but

results diverge and the methods used are very different –

ranging from direct measurements to the use of various

proxies – and Sidle et al. (2006) argue that most studies do

not distinguish between soil loss processes, e.g. land slides

vs. run-off. The mentioned studies often indicate that

shifting cultivation may be sustainable at low intensity, but

rarely specify this level in any detail (e.g. Juo and Manu,

1996). In a recent study, Mertz et al. (2008) found very poor

correlation between fallow length (and other intensity

indicators) and crop productivity and soil fertility in shifting

cultivation systems on Borneo, illustrating that the interac-

tions are not straightforward.

Lal (1987) summarized work at the International Institute

of Tropical Agriculture showing very low soil erosion of a

traditional farming practice resembling shifting cultivation

and studies in Thailand have shown sediments production

from shifting cultivation to be limited compared to other

land uses (Forsyth, 1994; Ziegler et al., 2004). In

Bangladesh, on the other hand, soil loss during run-off in

shifting cultivation was estimated to remove 27% of the

nutrients in the top 10 cm of the soil (Gafur et al., 2000),

though much of the soil material lost from the fields is

deposited within the watershed. Evidence from Guatemala

also indicates high erosion on shifting cultivation fields

(milpas) (Beach, 1998). Wairiu and Lal (2003) used soil

organic carbon (SOC) depletion as a proxy for soil erosion

and thus proposed strong erosion of shifting cultivation

compared to natural forest in the Solomon Islands. SOC

depletion following shifting cultivation has been observed in

other sites (van Noordwijk et al., 1997; Roder et al., 1997),

but was not observed in Sarawak where either no change in

SOC was observed (Bruun et al., 2006) or SOC levels

quickly recovered during fallow (Kendawang et al., 2004).

In Sarawak, recent studies have indicated the presence of

anthropogenic soil erosion as far back as 6000 years (Hunt

and Rushworth, 2005), but only one study have been carried

out on soil erosion of current land use practices. Using

experimental plots in two sites in Sarawak, Tanaka et al.

(2004) found that few nutrients were lost through run-off on

clayey soils and on sandy soils most nutrients were lost

through leaching, and in general impacts of shifting

cultivation on soil properties after more than 10 years of

fallow and 1 year of cultivation were found to be limited

(Tanaka et al., 2005; Kendawang et al., 2005).

The more intensive cultivation of black pepper (Piper

nigrum L.) often associated with shifting cultivation in

Sarawak has been even less studied in terms of run-off and

soil loss. In India, pure stands of black pepper resulted in

more soil erosion than cardamom or mixed stands of the two

(Moench, 1991) and in Sarawak, studies in the 1970s

indicated that black pepper cultivation caused higher erosion

than other upland farming practices (Hatch, 1981, 1982).

This is particularly important when black pepper is farmed

on clean-weeded, relatively steep slopes with poorly

constructed terraces, which are common in Sarawak

smallholder systems.

Soil texture, soil carbon, nutrients and 137Cs content –

emanating from fallout after nuclear bomb testing between

1952 and 1963 – has been widely used as indicators of soil

erosion (reviewed in e.g. Zapata, 2003). 137Cs is strongly

fixed in mineral complexes in clay soils, and hence

effectively labels the top soil at the time of deposition.

The mobility in soil from leaching and plant uptake is very

limited. For 137Cs, peak deposition occurred in 1963, thus137Cs is a suitable indicator for erosion studies on a 30–50

year timescale. The lower fallout on the southern hemi-

sphere and on equator compared to the northern hemisphere

makes the method more difficult to use there, however the

absence of later pollution with 137Cs from Chernobyl makes

calculations straightforward (Schuller et al., 2003).

Obviously, there are both gaps in knowledge and

considerable disagreement about the effects of smallholder

farming systems on soil erosion, and the aim of this paper is

to broaden the knowledge base by comparing soil erosion

and land degradation over the past 40 years in natural forest,

shifting cultivation of upland rice and black pepper in two

areas of Sarawak, Malaysia. The paper investigates the

impact of medium intensity shifting cultivation and more

intensive pepper cultivation on soil carbon inventories, soil

fertility and topsoil losses as estimated by 137Cs inventory.

2. Materials and methods

2.1. Study area

Samples were taken from the Niah Sub-District, Miri

Division and Padawan District, Serian Division in Sarawak,

Malaysia, both located within the lowland humid tropics with

average annual temperatures of 27 8C and little annual and

diurnal variation (Halenda, 1989). In Niah, annual precipita-

tion is 2700 mm; in Padawan about 4000 mm (Kuching

station). In both areas the driest month is August and the

wettest month is January. The soils sampled in both areas

resemble tropudults in the USDA Soil taxonomy (Tie et al.,

1989) (dystric nitosols according to FAO classification).

In May 2002 samples were collected in Niah from fields

and forest belonging to the Iban communities of Rumah

Muyang (42 households, location: 0384502400N and

11384505500E) and Rumah Ulat (63 households, location:

0380304600N and 11384404000E). Details of slope and relief of

samples fields are given in Table 1. In May 2003, samples

were collected in Padawan from fields and forest belonging to

the Bidayuh communities of Assom (30 households, location:

0181001800N and 11081300500E) and Parang (34 households,

location: 0181005200N and 11081205200E). The Padawan area

has steeper slopes than Niah, as described in Table 1.

The vegetation is characterised by tropical rainforest in

various stages of re-growth. The land use practices resemble

those described for other areas of Sarawak (Cramb, 1993;

184

Table 1

Characteristics of sampled fields

Cover Slope (8) Orientation of

slope (8)Years of cultivation Average fallow

length (years)

Years of

fallow

Padawan site

Rice 1 30–35 250 8 times since 1950 6.6

Rice 2 40 40 7 times since 1950 7.6

Rice 3 25 20 8 times since 1952 6.4

Pepper 1 30 40 8 years

Pepper 2 35 330 15 years

Pepper 3 35 80 22 years

Forest 1 30 300 >60

Forest 2 Variable, 20–50 0 >60

Forest 3 40 240 >60

Niah site

Rice 1 25 60 9 times since 1954 5.4

Rice 2 25 110 7 times since 1964 5.6

Rice 3 25 50 11 times since 1930 6.5

Pepper 1 20–25 180 7 years

Pepper 2 20–25 50 9 years

Pepper 3 – –

Forest 1 15–20 60 >50

Forest 2 15–20 340 60

Forest 3 25 60 >60

Mertz and Christensen, 1997; Wadley and Mertz, 2005):

fields for upland rice are cleared in June and usually burnt in

August, weather permitting. Upland rice (Oryza sativa L.) is

planted within 1 week of burning and harvested in February–

March. Fields are usually only cultivated 1 year, although

vegetables including chilli (Capsicum spp.), cassava (Man-

ihot esculenta Crantz), bananas (Musa spp.) and other

perennial crops may be cultivated and harvested several years

into the fallow period. Average fallow length in the Niah area

is about 11 years and in Padawan about 13 years (Mertz et al.,

2008), hence the fields selected for sampling are more

intensively cultivated than the average in the area (Table 1).

Field sizes vary greatly, but average about 1 ha. Pepper (Piper

nigrum L.) is cultivated on separate, permanent fields that are

maintained for up to 22 years, but averaging 8 years. High

infection levels of pests and diseases are common reasons for

abandoning an area and establishing new fields. Undisturbed

forest plots were identified by local farmers and included

sacred forest, protected areas and fields that had not been

cultivated for more than 60 years.

2.2. Sampling for nutrient content and soil quality

In Niah, soil samples from 16 upland rice fields were

collected with a soil core auger (Eijkelkamp, i.d. 34 mm) to

30 cm depth. 25 soil cores taken within a 10 m � 10 m

square were combined in one sample. Samples were

collected from top of slope, middle of slope and bottom

of slope, three at each level, totalling nine independent

samples from each plot (each consisting of 25 cores). Soil

was air dried, crushed and mixed and sub-samples were

taken for analysis at the laboratory of Semongok Agricul-

tural Research Centre, Kuching, Sarawak.

2.3. Sampling for 137Cs and soil carbon

In Niah, soil samples were collected from three upland

rice fields, two pepper gardens and three undisturbed forest

sites. Soil was collected to 90 cm depth with a soil core

auger (Eijkelkamp, i.d. 18 mm), and divided in 0–30 cm,

30–60 cm and 60–90 cm intervals. 6 cores taken within a

10 m � 10 m square were combined in one sample. Three

samples (each consisting of 6 cores) were taken upslope,

three midslope and three downslope, totalling nine

independent samples for each plot. The samples were air

dried and transported to Copenhagen, Denmark for analysis.

In Padawan, soil samples were collected from three

upland rice fields, three pepper gardens and three

undisturbed forest sites. Soil was collected volume specific

with soil sample rings (Eijkelkamp, 5 cm internal diameter;

100 cm3) from the vertical face of exposed soil profiles.

Samples were collected at 0–5 cm, 15–20 cm, 40–45 cm and

70–75 cm depth, representing the following intervals:

0–5 cm, 5–30 cm, 30–60 cm and 60–90 cm.

Samples were collected upslope, midslope and down-

slope. At each level, duplicate soil ring cores from three

adjacent soil profiles were collected and combined to one

sample (consisting of 6 rings). Three samples were collected

from each level (upslope, midslope and downslope),

totalling 9 independent samples. The samples were air

dried and transported to Copenhagen, Denmark for analysis.

2.4. Sample analyses

Soil samples were analysed at the Chemistry Laboratory

at the Agricultural Research Centre, Semongok in Sarawak.

Particle size distribution was determined by the Hydrometer

Fig. 1. Average pH and extractable P and Ca in topsoil as function of

position on slope from 16 upland rice fields in Niah, Sarawak. Average

values from 6 level fields are indicated separately. Error bars indicate S.E.

(n = 16). pH and Ca are significantly affected by position on slope

(P < 0.001 and P < 0.05, respectively).

Method (FAO, 1970), contents of soil organic carbon by

the Walkley–Black Method (Nelson and Sommers, 1996),

total nitrogen by the Dumas combustion method (Bremner,

1996), available phosphorus by the Bray and Kurtz (1945)

method, Cation Exchange Capacity (CEC) by the

Ammonium Acetate Method buffered at pH 7 (Sumner

and Miller, 1996), and contents of exchangeable basic

cations by leaching the soil with 1 M ammonium acetate

buffered to pH 7 (Thomas, 1982). Subsequently the

contents of cations in the leachate were measured by

means of AAS. pH was measured at a 1:2.5 soil:water ratio

at a pH meter with an accuracy better than 0.05 unit (Chin,

2000).

Soil bulk density was determined by drying the soil ring

samples (100 cm3) at 105 8C and weighing them. In

Denmark, samples were finely ground and analysed for

total C and N content using an mass spectrometer (Europa

Scientific, 20–20) coupled to an ANCA-SL sample

preparation module (Europa Scientific, Crewe, UK). Soil

C mass was calculated using bulk density and soil carbon

concentration.137Cs content was measured at the Danish National

Institute of Radiation Protection, Copenhagen using a

sample of approximately 200 g sieved soil in cylindrical

containers 5 cm high, 8 cm diameter. The soil content of137Cs was measured around 662 keV on a cooled Germa-

nium spectrometer coupled to a multi-channel analyser

using GammaView software. Due to the low activity of the

samples, counting time was 24 h, yielding a lower detection

limit of 0.3 Bq kg�1.

2.5. Statistical analysis

The experiment compared three land use types (shifting

cultivation rice, forest and pepper), three positions on slope

(top, middle and bottom), and three (Niah) or four

(Padawan) soil depths. The data were statistically analysed

by ANOVA. Differences between land uses, slope position

and soil depths were considered significant when LSD

values had P values lower than 0.05.

Fig. 2. Extractable Mg and K in topsoil as function of position on slope

from 16 upland rice fields in Niah, Sarawak. Average values from 6 level

fields are indicated separately. Error bars indicate S.E. (n = 16). Both are

significantly affected by position on slope (P < 0.001).

3. Results

3.1. Niah data

Soil pH, extractable Ca and P from upland rice fields in

Niah are depicted in Fig. 1. Soil pH and extractable Ca were

significantly affected by position on slope (P < 0.001 and

P < 0.05, respectively). Highest values were found at the

bottom of the slope, lowest values at the top. Level fields

showed intermediate values. Extractable soil P showed the

same trend, but differences between sample positions on the

slope were not significant.

Soil extractable Mg and K from the same fields

demonstrated same trend as Ca and pH (Fig. 2). Values

were significantly lower at top of slope than bottom of slope

(P < 0.001 for both Mg and K).

Soil solute conductivity in shifting cultivation rice fields

showed same trend as the nutrients in Figs. 1 and 2, namely a

significant effect of sample position on the slope and a

higher level at the bottom of the slopes (P < 0.001) (Fig. 3).

All other measured parameters than those depicted in

Figs. 1–3 (total soil C and N, extractable Na, CEC and soil

texture), did not show any significant correlation with

position on slope.137Cs activity (Bq kg�1 soil) in three depths of the soils

from upland rice fields, pepper gardens and undisturbed

forest are shown in Fig. 4. The activity was not significantly

affected by land use in the top soil layer, but significantly

higher in the 30–60 cm layer in undisturbed forest than in the

two cultivated areas.

3.2. Padawan data

Soil carbon concentration (in %) in upland rice fields,

pepper gardens and undisturbed forest at four soil depth

intervals are shown in Fig. 5. Carbon content was

significantly higher in forest and rice than in pepper in

Fig. 3. Topsoil conductivity as function of position on slope from 16 upland

rice fields in Niah, Sarawak. Average values from 6 level fields are indicated

separately. Error bars indicate S.E. (n = 16) (P < 0.001).

Fig. 5. Soil organic carbon concentration in four soil layers from upland

rice fields, pepper gardens and native forest in Padawan, Sarawak. Carbon

concentration is significantly lower in the top 5 cm of pepper gardens

compared to the other land uses (P < 0.001). Error bars show S.E. (n = 3,

each consisting of 9 independent samples).

the top 5 cm of soil (P < 0.001). For deeper soil layers, there

was no effect of land use on C content.

Total carbon inventory in kg C m�2 to 90 cm depth in

upland rice, pepper gardens and undisturbed forest are

shown in Fig. 6. Carbon inventory was calculated on the

basis of carbon content and soil density at the different soil

depths. Soil carbon stocks were not significantly affected by

land use in any of the soil depths, nor in the profile as a

whole. There was also no significant effect of position on

slope on the soil carbon stocks.137Cs concentration in the soil (Bq kg�1 soil) at four depth

intervals in shifting cultivation rice, pepper gardens and

undisturbed forest is shown in Fig. 7. 137Cs concentration in

topsoil was significantly affected by land use (P < 0.001);

undisturbed forest contained four times as much 137Cs as the

pepper gardens and twice as much as the upland rice fields.

There were no significant differences in 137Cs in the subsoil

layers, forest subsoil below 30 cm contained less 137Cs than

the detection limit. Total 137Cs inventory in Bq m�2 to 90 cm

depth is shown in Fig. 8. Most subsoil samples in all land uses

were close to or below the detection limit (average values

0.15–0.4 Bq kg�1, detection limit 0.3 Bq kg�1), hence the

uncertainty of the estimates of subsoil 137Cs inventory is high.

Even with just a few samples above the detection limit, the

Fig. 4. 137Cs activity in three soil layers from upland rice fields, pepper

gardens and native forest plots in Niah, Sarawak. Error bars indicate S.E.

(n = 3, each consisting of 9 independent samples).

137Cs inventory in forest subsoil would have been comparable

to the other land use types. Consequently, in the discussion,

only the top soil layer will be evaluated. There was no

significant effect of position on slope (top, middle, bottom) on

the 137Cs content of the soil.

4. Discussion

As described in the introduction, the effects of shifting

cultivation and other smallholder farming practices on soil

erosion are not well understood. Empirical data sets are in

disagreement – both in data shown and interpretation – and

often text books and other literature uses anecdotal rather

than strong empirical data for describing shifting cultivation

as environmentally detrimental (Watson, 1989; Beets, 1990;

Rasul and Thapa, 2003). This study shows that in the case of

shifting cultivation practices in Niah and Padawan Districts

of Sarawak, Malaysia, there is indeed a limited amount of

soil erosion from shifting cultivation, but the magnitude of

this erosion must be analysed in the context of alternative

systems.

Fig. 6. Total organic carbon in profile, calculated from carbon concentra-

tion and bulk density of soil, from upland rice fields, pepper gardens and

native forest in Padawan, Sarawak. Land uses are not significantly different.

Error bars show S.E. (n = 3, each consisting of 9 independent samples).

Fig. 7. 137Cs activity in four soil layers from upland rice fields, pepper

gardens and native forest plots in Padawan, Sarawak. 137Cs content in

subsoil of upland rice fields was below detection limit (0.3 Bq kg�1 soil).

Error bars indicate S.E. (n = 3, each consisting of 9 independent samples).

4.1. Soil chemical indicators

In the present study, the soil content of base-forming

cations, K, Ca, Mg and Na, showed significantly higher

values on the downslope samples compared to upslope

samples. Similarly, conductivity was significantly greater in

the downslope samples. As soil conductivity is an integral of

all dissolved ions, and Ca and Mg make up a significant part

of the total cations, it is not surprising that conductivity

exhibits the same trend as these ions.

These results indicate a larger loss of cations from the

upslope soil, but are not necessarily an indication of soil

erosion. The upslope soils are better drained than the

downslope areas, which are often close to a river or wetland.

Downward water movement will eventually leach out

cations from the soil, and this process will be most

pronounced in the upslope areas. Another possible

mechanism is the loss of ash by wind or water transport

immediately after burning the forest. Ashes contain the non-

volatile minerals from the burnt plant material, including the

base-forming cations. It is likely that there can be a

downward movement of ashes in the period between burning

Fig. 8. Total 137Cs in soil profile, calculated from 137Cs activity and bulk

density of soil, from upland rice fields, pepper gardens and native forest

plots in Padawan, Sarawak. 137Cs content in subsoil of upland rice fields was

below detection limit (0.3 Bq kg�1 soil). Error bars indicate S.E. (n = 3,

each consisting of 9 independent samples).

and stabilisation of the ash layer by emerging rice plants

(Bruun et al., 2006). This would also explain why it was only

the non-volatile cations found in the ashes (Ca, K, Mg), and

parameters affected by these (conductivity) increased at the

foot of the slope, while other measured soil parameters that

would mainly change as a direct effect of soil movement

(soil carbon, texture and CEC) were not affected.

The lower pH upslope can also be explained by the loss of

cations with percolating water or from ashes. As the base-

forming cations leach out, the relative dominance of H+ ions

of the exchangeable cations and cations in solution

increases, leading to a decrease in pH. Extractable soil P

showed the same trend as the cations and pH, but the

difference between slope positions was not significant.

Phosphorus is much less soluble in soil solution than the

cations, and only leaches from soils in significant amounts

when the sorption capacity of the soils is exceeded, which is

far from the case in the study area. The less pronounced

effect on soil P would therefore indicate that downhill

transport of topsoil is less significant in explaining the

differences between upslope and downslope samples, as this

mechanism should effect the cations and phosphorus in the

same way.

None of the other measured soil quality indicators (clay

content, cation exchange capacity, organic carbon and

nitrogen) were significantly different between slope posi-

tions. Loss of any of these parameters (and partly P as well),

is usually associated with loss of topsoil by soil movement—

erosion, either downward by micro-erosion through soil

cracks and pores or downslope during run-off. The fact than

none of these parameters are depleted in the upslope

samples, indicates that erosion from these sites and

subsequent deposition at the foot of the hill was not a

significant process in this area. In a similar study researchers

found no steep gradients in soil pH and phosphorus across

slopes and concluded that crops and residue formed a natural

barrier against soil erosion during cultivation (Rodenburg

et al., 2003).

4.2. Soil carbon

Analysis of carbon in the soil profile from the Padawan

site highlighted the importance of sampling the entire profile

and correcting for soil density. In the top soil layer, pepper

gardens contained significantly lower concentrations (C g�1

soil) of carbon, than the rice and forest plots. However, due

to significantly higher soil density in the pepper gardens

(data not shown), the total amount of carbon (g m�2) was

equal in all three land uses. In the soil layers below 5 cm

depth, both density and carbon concentration were similar in

all three land uses. The contribution to the total carbon

inventory from the top 5 cm of the profile (where both

density and carbon concentration were affected by land use)

is marginal, constituting less than 10% for all three land

uses. Hence, even if total carbon content of the top of the

profile would be affected by land use, this is not likely to

188

affect the carbon storage of the profile as a whole. However,

top soil properties may be significantly affected by the

variation in carbon content and bulk density, affecting water

infiltration rates, nutrient dynamics and root interception.

The results highlight the importance of volume specific

sampling within the entire soil profile when evaluating land

use effects on soil quality.

Soils were sampled to 90 cm depth. It was not suspected

that land use would affect soil C content below this depth.

The finding that only the top 5 cm were significantly affected

supports this assumption; hence sampling to greater depth is

unlikely to have changed the result.

Other researchers have reported similar findings.

Yemefack et al. (2006) reported from a study of shifting

cultivation in Cameroon that soil C was much less affected

than other indicators as pH, Ca and P when sampled to

20 cm. They explained this by a ‘‘dilution’’ effect of the Ah

horizon and top soil layers by deeper (<10 cm) soil layers.

4.3. Soil 137Cs

In the Niah site, 137Cs activity in the topsoil layer was not

affected by land use/cover, but in the 30–60 cm depth

interval, the activity was significantly higher in the

undisturbed forest. The soils were sampled using a 90 cm

soil core auger (18 mm inner diameter), and then dividing

the soil tube in the relevant depth intervals. The high content

of 137Cs in the deeper soil layers, together with observations

during sampling, led to suspicion that contamination from

the topsoil to the deeper layers occurred during sampling. As137Cs is very immobile in soils with active clays, it is

difficult to imagine which processes would lead to an

accumulation in the subsoil of an undisturbed forest.

However, micro-erosion of clay particles through macro-

pores in the soil could be such a pathway. In order to avoid

the possibility of contamination of subsoils during sample

collection, the sampling strategy was changed at the

Padawan site, also allowing for calculation of the total Cs

and C inventory by taking volume specific samples.

At the Padawan site, the separate sampling of the 0–5 cm

layer, yielded much higher activities (>8 Bq kg�1 soil) than

the sampling of the 0–30 cm depth interval at Niah. This

illustrates how Cs is confined to the top soil layers, if physical

movement of soil has not altered this distribution. The forest

sites contained about four times the activity of the pepper

gardens, and about twice the activity of the rice fields.

When corrected for bulk density, the differences are less

pronounced. If excluding the data for the 5–30 cm layer (due

to uncertainty of estimates in rice and pepper fields caused

by the low activity of 137Cs in the soils), the Cs inventory in

rice and pepper fields is 82 and 65% of the forest soil,

respectively, indicating soil losses of 18 and 35%, but only

from a very shallow soil layer. If including the 5–30 cm

layer, the Cs inventory is just around 70% for both land uses.

The combination of steep slopes (15–508 in this study),

clearing of land and exposing the soil, combined with heavy

rainfall events and an annual precipitation of 2500–

4000 mm, certainly can be conductive for significant soil

erosion. Studies have shown a relatively direct correlation

between slope gradient and erosion (Zhang et al., 2003). In

other tropical farming systems, with steep slopes and high

rainfall intensity, traditional farming systems have devel-

oped well functioning erosion control measures (Malley

et al., 2004; He et al., 2007). The two communities in

Sarawak do not at all consider soil erosion an issue, and

mainly associate the term with rare incidents of landslides,

indicating that soil erosion is not obvious at field level. The

shifting cultivation system practiced by Iban and Bidayuh

communities in Sarawak is characterised by a quite short

period of exposed soil. After cutting the trees, the soil is still

covered by undergrowth, besides the branches and falling

leaves. After burning, litter often remains in the fields due to

incomplete combustion, and vegetables and upland rice are

planted within a few days using a planting stick—hence the

soil disturbance is minimal. The system could rightfully be

compared to conservation tillage systems, which are

recommended for their erosion controlling abilities. The

clearing period also coincides with the driest months of the

year, in order to facilitate burning of the vegetation; hence

the impact of rainwater is minimal. After planting, the rice

and vegetables (and weeds) form a closed canopy within 3–6

weeks. The mechanical weeding that takes place has in other

studies been shown to cause only minor soil erosion, when

compared to the water erosion in the system (Turkelboom

et al., 1999; Ziegler et al., 2007). The main reason for this is

the very shallow ‘‘tillage’’ depth of 1–2 cm. During harvest

only the rice panicle is removed, leaving the straw, residual

crops, weeds and undergrowth to develop into fallow

vegetation with complete soil coverage already from the

time of harvest. Similar studies have highlighted the

importance of un-burnt material, crops and crop residue

in controlling surface erosion (Rodenburg et al., 2003).

Consequently, the soil is only exposed for a period of 4–8

weeks during the driest months of each cropping cycle and

with fallow lengths of 5–10 years, this only result in an

average annual soil exposure of about 1 week. Hence, it is

not surprising that we find the cumulated soil loss over the

past 40 years in areas affected by shifting cultivation to be

very limited. In a short review, Juo and Manu (1996)

presented the view that as long as shifting cultivation fields

were small and surrounded by forest, the cultivation period

<2 years, and the fallow period sufficiently long (not

specified), shifting cultivation may adequately conserve soil

nutrient stocks. However, they also expressed the view, that

this situation was very rare, and increasingly so. This study

suggests that these ideal conditions may not be so rare.

5. Conclusions

When evaluating the environmental effects of shifting

cultivation, it must be compared with alternative farming

practices—not only with forestry or natural forest. Being a

farming system that relies on field crop production, shifting

cultivation cannot be expected to be more environmentally

friendly than natural forest, but it would in most cases provide

more and better environmental services than other more

intensive farming systems. The forest and crop litter,

combined with the minimal soil disturbance in the shifting

cultivation systems, forms an almost continuous covering

layer that protects the soil from erosion. Although we

observed a gradient in selected soil nutrients down the slopes,

this is more likely attributed to ash dispersal after burning,

rather than soil erosion. The carbon inventory was not affected

by land use, and although the topsoil of the cultivated land was

depleted of 137Cs compared to the natural forest, the total soil

loss after numerous cultivation cycles was calculated to be

<30%. We conclude that at the given cultivation intensity, soil

fertility and quality loss as a result of shifting cultivation is of

minor importance, and besides the temporary loss of above

ground biomass, these systems are not net emitters of carbon

compared to natural forest.

Acknowledgements

The research was funded by the Danish SLUSE

university network for Sustainable Land Use and Natural

Resource Management. The authors are indebted to Klaus

Ennow from Danish National Institute for Radiation

Protection for handling the Cs analysis, and Jona Anak

Kerani for field assistance. We thank the villagers from

Rumah Muyang and Kampung Assom for their hospitality

and kind cooperation.

References

Beach, T., 1998. Soil catenas, tropical deforestation, and ancient and

contemporary soil erosion in the Peten, Guatemala. Phys. Geogr. 19,

378–405.

Beets, W.C., 1990. Raising and Sustaining Productivity of Smallholder

Farming Systems in the Tropics. AgBe Publishing, Alkmaar, The

Netherlands.

Borggaard, O.K., Gafur, A., Petersen, L., 2003. Sustainability appraisal of

shifting cultivation in the Chittagong Hill Tracts of Bangladesh. Ambio

32, 118–123.

Brady, N.C., 1996. Alternatives to slash-and-burn: a global imperative.

Agric. Ecosyst. Environ. 58, 3–11.

Bray, R.H., Kurtz, L.T., 1945. Determination of total, organic, and available

forms of phosphorus in soils. Soil Sci. 59, 39–45.

Bremner, J.M., 1996. Nitrogen—total. In: Sparks, D.L., Page, A.L.,

Helmke, P.A., Loeppert, R.H., Soltantpour, P.N., Tabatabi, M.A.,

Johnston, C.T., Sumner, M.E. (Eds.), Methods of Soil Analysis, Part

3—Chemical Methods. Soil Science Society of America, Madison,

pp. 1087–1089.

Bruun, T.B., Mertz, O., Elberling, B., 2006. Linking yields of upland rice in

shifting cultivation to fallow lenght and soil properties. Agric. Ecosyst.

Environ. 113, 139–149.

Chin, S.P., 2000. A Laboratory Manual of Methods of Soil Analysis.

Department of Agriculture Sarawak, Kuching.

Cramb, R.A., 1993. Shifting cultivation and sustainable agriculture in East

Malaysia: a longitudinal case study. Agric. Syst. 42, 209–226.

de Jong, W., 1997. Developing Swidden agriculture and the threat of

biodiversity loss. Agric. Ecosyst. Environ. 62, 187–197.

Devendra, C., Thomas, D., 2002. Smallholder farming systems in Asia.

Agric. Syst. 71, 17–25.

FAO, 1970. Physical and Chemical Methods of Soil and Water Analysis.

Food and Agriculture Organization, Rome.

Forsyth, T.J., 1994. The use of Cesium-137 measurements of soil erosion

and farmer’s perceptions to indicate land degradation amongst shifting

cultivators in northern Thailand. Mountain Res. Dev. 14, 229–244.

Fox, J., 2000. How blaming ‘slash and burn’ farmers is deforesting mainland

Southeast Asia. Asia Pacific Iss. 47, 1–8.

Gafur, A., Borggaard, O.K., Jensen, J.R., Petersen, L., 2000. Changes in soil

nutrient content under shifting cultivation in the Chittagong Hill Tracts

of Bangladesh. Geografisk Tidsskrift, Danish J. Geogr. 100, 37–46.

Halenda, C.J., 1989. The ecology of fallow forest after shifting cultivation in

Niah Forest Reserve. Forest Research Report. Forest Department,

Kuching, Malaysia.

Harwood, R.R., 1996. Development pathways toward sustainable systems

following slash-and-burn. Agric. Ecosyst. Environ. 58, 75–86.

Hatch, T., 1981. Preliminary results of soil erosion and conservation trials

under pepper (piper nigrum) in Sarawak, Malaysia. In: Morgan, R.P.C.

(Ed.), Soil Conservation Problems and Prospects. John Wiley & Sons,

Chichester, pp. 255–262.

Hatch, T., 1982. Shifting Cultivation in Sarawak: A Review. Soils Division,

Research Branch, Department of Agriculture, Kuching, Sarawak.

He, X., Xu, Y., Zhang, X., 2007. Traditional farming system for soil

conservation on slope farmland in southwestern China. Soil Tillage

Res. 94, 193–200.

Hunt, C.O., Rushworth, G., 2005. Cultivation and human impact at 6000 cal

yr BP in tropical lowland forest at Niah, Sarawak, Malaysian Borneo.

Quat. Res. 64, 460–468.

Juo, A.S.R., Manu, A., 1996. Chemical dynamics in slash-and-burn agri-

culture. Agric. Ecosyst. Environ. 58, 49–60.

Kendawang, J.J., Tanaka, S., Ishihara, J., Shibata, K., Sabang, J., Ninomiya,

I., Ishizuka, S., Sakurai, K., 2004. Effects of shifting cultivation on soil

ecosystems in Sarawak, Malaysia I. Slash and burning at Balai Ringin

and Sabal experimental sites and effect on soil organic matter. Soil Sci.

Plant Nutr. 50, 677–687.

Kendawang, J.J., Tanaka, S., Shibata, K., Yoshida, N., Sabang, J., Ninomiya,

I., Sakurai, K., 2005. Effects of shifting cultivation on soil ecosystems in

Sarawak, Malaysia. III. Results of burning practice and changes in soil

organic matter at Niah and Bakam experimental sites. Soil Sci. Plant

Nutr. 51, 515–523.

Kleinman, P.J.A., Pimentel, D., Bryant, R.B., 1995. The ecological sustain-

ability of slash-and-burn agriculture. Agric. Ecosyst. Environ. 52, 235–

249.

Lal, R., 1974. Soil erosion and shifting agriculture. In: FAO. (Eds.), Shifting

Cultivation and Soil Conservation in Africa. FAO, Rome, pp. 49–71.

Lal, R., 1987. Need for, approaches to and consequences of land clearing

and development in the tropics. In: IBSRAM (Ed.), Tropical Land

Clearing for Sustainable Agriculture. Proceedings of an IBSRAM

Inaugural Workshop held in Jakarta and Bukittingi, Indonesia 27/8–

3/9, 1985. IBSRAM, Bangkok, pp. 15–27.

Malley, Z.J.U., Kayombo, B., Willcocks, T.J., Mtakwa, P.W., 2004. Ngoro: an

indigenous, sustainable and profitable soil, water and nutrient conserva-

tion system in Tanzania for sloping land. Soil Tillage Res. 77, 47–58.

Mertz, O., 2002. The relationship between fallow length and crop yields in

shifting cultivation: a rethinking. Agrofor. Syst. 55, 149–159.

Mertz, O., Christensen, H., 1997. Land use and crop diversity in two Iban

communities, Sarawak, Malaysia. Geografisk Tidsskrift, Danish J.

Geogr. 97, 98–110.

Mertz, O., Wadley, R.L., Nielsen, U., Bruun, T.B., Colfer, C.J.P., de

Neergaard, A., Jepsen, M.R., Martinussen, T., Zhao, Q., Noweg,

G.T., Magid, J., 2008. A fresh look at shifting cultivation: fallow length

an uncertain indicator of productivity. Agric. Syst. 96, 75–84.

Moench, M., 1991. Soil-erosion under a successional agroforestry

sequence—a case-study from Idukki District, Kerala, India. Agrofor.

Syst. 15, 31–50.

Nelson, D.W., Sommers, L.E., 1996. Total carbon, organic carbon and

organic matter. In: Sparks, D.L., Page, A.L., Helmke, P.A., Loeppert,

R.H., Soltantpour, P.N., Tabatabi, M.A., Johnston, C.T., Sumner, M.E.

(Eds.), Methods of Soil Analysis, Part 3—Chemical Methods. Soil

Science Society of America, Madison, pp. 995–996.

Pimentel, D., Harvey, D., Resosudarmo, P., Sinclair, K., Kurz, D., McNair,

M., Crist, S., Shpritz, L., Fitton, L., Saffouri, R., Blair, R., 1995.

Environmental and economic costs of soil erosion and conservation

benefits. Science 267, 1117–1123.

Pimentel, D., 2006. Soil erosion: a food and environmental threat. Environ.

Dev. Syst. 8, 119–137.

Rasul, G., Thapa, G.B., 2003. Shifting cultivation in the mountains of South

and Southeast Asia: regional patterns and factors influencing the

change. Land Degrad. Dev. 14, 495–508.

Rasul, G., Thapa, G.B., Zoebisch, M.A., 2004. Determinants of land-use

changes in the Chittagong Hill Tracts of Bangladesh. Appl. Geogr. 24,

217–240.

Rodenburg, J., Stein, A., van Noordwijk, M., Ketterings, Q.M., 2003.

Spatial variability of soil pH and phosphorus in relation to soil run-

off following slash-and-burn land clearing in Sumatra, Indonesia. Soil

Tillage Res. 71, 1–14.

Roder, W., Phengchanh, S., Maniphone, S., 1997. Dynamics of soil and

vegetation during crop and fallow period in slash-and-burn fields of

northern Laos. Geoderma 76, 131–144.

Schmidt-Vogt, D., 1998. Defining degradation: the impacts of swidden

on forests in northern Thailand. Mountain Res. Dev. 18, 135–

149.

Schuller, P., Ellies, A., Castillo, A., Salazar, I., 2003. Use of 137Cs to

estimate tillage- and water-induced soil redistribution rates on agricul-

tural land under different use and management in central-south Chile.

Soil Tillage Res. 69, 69–83.

Sidle, R.C., Ziegler, A.D., Negishi, J.N., Nik, A.R., Siew, R., Turkelboom,

F., 2006. Erosion processes in steep terrain - Truths, myths, and

uncertainties related to forest management in Southeast Asia. For. Ecol.

Man. 224, 199–225.

Sumner, M.E., Miller, W.P., 1996. Cation exchange capacity and exchange

coefficients. In: Sparks, D.L., Page, A.L., Helmke, P.A., Loeppert,

R.H., Soltantpour, P.N., Tabatabi, M.A., Johnston, C.T., Sumner,

M.E. (Eds.), Methods of Soil Analysis, Part 3—Chemical Methods.

Soil Science Society of America, Madison, pp. 1220–1221.

Tanaka, S., Kendawang, J.J., Ishihara, J., Shibata, K., Kou, A., Jee, A.,

Ninomiya, I., Sakurai, K., 2004. Effects of shifting cultivation on soil

ecosystems in Sarawak, Malaysia II. Changes in soil chemical proper-

ties and runoff water at Balai Ringin and Sabal experimental sites. Soil

Sci. Plant Nutr. 50, 689–699.

Tanaka, S., Kendawang, J.J., Yoshida, N., Shibata, K., Jee, A., Tanaka, K.,

Ninomiya, I., Sakurai, K., 2005. Effects of shifting cultivation on soil

ecosystems in Sarawak, Malaysia—IV. Chemical properties of the soils

and runoff water at Niah and Bakam experimental sites. Soil Sci. Plant

Nutr. 51, 525–533.

Thomas, G.W., 1982. Exchangeable Cations. In: Page, A.L., Miller, R.H.,

Keeney, D.R. (Eds.), Methods of Soil Analysis, Part 2, Chemical and

Microbiological Properties. American Society of Agronomy and Soil,

Science Society of America, Madison, WI.

Tie, Y.L., Ng, T.T., Chai, C.C., 1989. Hill padi-based cropping system in

Sarawak, Malaysia. In: van der Heide, J. (Ed.), Nutrient Management

for Food Crop Production in Tropical Farming Systems. Institute for

Soil Fertility, Haren, The Netherlands, pp. 301–312.

Turkelboom, F., Poesen, J., Ohler, I., Ongprasert, S., 1999. Reassessment of

tillage erosion rates by manual tillage on steep slopes in northern

Thailand. Soil Tillage Res. 51, 245–259.

van Noordwijk, M., Cerri, C., Woomer, P.L., Nugroho, K., Bernoux, M.,

1997. Soil carbon dynamics in the humid tropical forest zone. Geoderma

79, 187–225.

Wadley, R.L., Mertz, O., 2005. Pepper in a time of crisis: smallholder

buffering strategies in Sarawak, Malaysia and West Kalimantan, Indo-

nesia. Agric. Syst. 85, 289–305.

Wairiu, M., Lal, R., 2003. Soil organic carbon in relation to cultivation and

topsoil removal on sloping lands of Kolombangara, Solomon Islands.

Soil Tillage Res. 70, 19–27.

Watson, J.W., 1989. The evolution of appropriate resource–management

systems. In: Berkes, F. (Ed.), Common Property Resources. Pinter,

Belhaven, pp. 55–69.

Yemefack, M., Jetten, V.G., Rossiter, D.G., 2006. Developing a minimum

data set for characterizing soil dynamics in shifting cultivation systems.

Soil Tillage Res. 86, 84–98.

Zhang, X., Zhang, Y., Wen, A., Feng, M., 2003. Assessment of soil losses on

cultivated land by using the 137Cs technique in the Upper Yangtze River

Basin of China. Soil Tillage Res. 69, 99–106.

Zapata, F., 2003. The use of environmental radionuclides as tracers in soil

erosion and sedimentation investigations: recent advances and future

developments. Soil Tillage Res. 69, 3–13.

Ziegler, A.D., Giambelluca, T.W., Sutherland, R.A., Nullet, M.A., Yarnasarn,

S., Pinthong, J., Preechapanya, P., Jaiaree, S., 2004. Toward under-

standing the cumulative impacts of roads in upland agricultural water-

sheds of northern Thailand. Agric. Ecosyst. Environ. 104, 145–158.

Ziegler, A.D., Giambelluca, T.W., Sutherland, R.A., Nullet, M.A., Vien,

T.D., 2007. Soil translocation by weeding on steep-slope swidden fields

in northern Vietnam. Soil Tillage Res. 96, 219–233.