Burkina Faso: Using Termites to Rehabilitate Degraded Soil in the Sahel


Like many places in the Sahel, Bam Province in northern Burkina Faso has suffered severe soil degradation as a result of intensive agriculture and overgrazing. Degraded areas develop a hard crust on the soil’s surface that reduces infiltration capacity and leads to increased runoff of rainfall. The soil’s inability to absorb needed moisture precipitates nutrient imbalances that compromise productivity and ultimately decrease per capita food production. In order to address this growing concern and evaluate different strategies for rehabilitating crusted soils, an experiment was conducted using varying combinations of termites and mulch to stimulate natural regenerative processes. By delineating termite and nontermite experimental plots that either incorporated mulch or were left bare, researchers were able to identify specific factors responsible for favorable signs of regeneration. The study showed that plots with termite activity had greater vegetation cover than sites without; and sites incorporating both termites and mulch were the most productive of all. The lessons afforded by this research have important implications for land management and rehabilitation in the region and, ultimately, for the livelihoods of local communities.

Quick Facts

Project Location:
Burkina Faso, 14.1000865, -0.14949879999994664

Geographic Region:

Country or Territory:
Burkina Faso

Desert/Arid Land


Organization Type:
University / Academic Institution


Project Stage:

Start Date:

End Date:

Primary Causes of Degradation

Agriculture & Livestock, Urbanization, Transportation & Industry

Degradation Description

In the Sahelian zone of West Africa, the combined effect of soil organic matter depletion due to overgrazing, continuous cultivation, and the decrease in rainfall in the last 30 years has resulted in the increase of crusted and unproductive land. Such land is characterized by low infiltration capacity, nutrient imbalance, reduced biodiversity, and very low to zero primary production. Annual runoff resulting from these degraded soils is estimated at 60-80% of annual rainfall on Cambisols and 50-90% on Lixisols.

Reference Ecosystem Description

The vegetation is of the steppe type according to UNESCO’s (1973) classification, consisting mostly of annual herbs and shrubs with few annual grasses and large bare areas. The life form for the grasses is dominantly annual therophytes; perennial grasses are rare (Mando 1991). The woody component is dominated by shrubs. The most important families are Mimosaceae and Combretaceae. The common termite genera are Microtermes, Macrotermes, and Odontotermes.

Lixisols and Cambisols are the most common soil types in the area (Burreau National des Sols 1995). The soil types on the experimental site are chromic cambisol, ferric lixisol, and haplic lixisol.

Chromic Cambisol has a pH between 5 and 8 from 5 cm to 120 cm depth, cation exchange capacity (CEC) < 17 cmol (+)/kg from 0 cm to 120 cm depth. The textural classes according to the U.S. Department of Agriculture system are sandy clay loam in the top 5 cm (49% sand, 29% silt, and 21% clay) and clay below 70 cm depth (45% clay, 20% sand, and 35% silt). Average bulk density is 1.6 g/cm3 at 0-5 cm depth and 1.7 g/cm3 at 120 cm depth. Soil organic matter content is 0.6% between 0 cm and 20 cm depth and 0.2% at 120 cm depth.

Ferric Lixisol has a pH of about 4.1 at 5 cm depth and 6.1 at 110 cm depth. CEC is 6 cmol(+)/kg at 5 cm depth and 12.4 cmol(+)/kg at 120 cm depth. The textural class is sandy loam (20% clay, 43% silt, and 37% sand). The average bulk density is 1.5 g/cm3 at 5 cm depth and 1.7 g/cm3 at depths below 70 cm. Soil organic matter content is 0.51% at 5 cm depth but decreases to 0.17% at 120 cm depth.

Haplic Lixisol has pH < 5, CEC < 9 cmol(+)/kg, and sandy loam textural class (53% sand, 25% silt, and 21% clay). Average bulk density is 1.4 g/cm3 at 5 cm depth and 1.6 g/cm3 at 120 cm depth. Soil organic matter content is 0.3% at 5 cm depth and 0.17% below 100 cm depth.

Project Goals

Termites are a major component of the soil fauna in the tropics (Lee & Wood 1971), and their importance in modifying soil properties is generally recognized (Lobry de Bruyn & Conacher 1990; Lavelle et al. 1992). When established on degraded soil, termites can improve soil physical properties within a short time (Mando 1997). They also play a key role in the decomposition process (Whitford et al. 1991). Yet, despite the relative abundance of information on termite ecology, little is known of the effect of termite-mediated processes on ecosystem productivity. Thus, the aim of this project was to evaluate the extent to which termite-mediated processes in crusted soil can rehabilitate the vegetation.


The project does not have a monitoring plan.

Description of Project Activities:
The experiment was laid out on a wasteland of completely unvegetated ground. The site was fenced to exclude large herbivores, and three blocks were delineated--one on each soil type--in a split-plot design. The main treatment was an insecticide, to obtain termite-infested plots (termite plots) and nontermite plots. We used Dieldrin at a rate of 500 g a.i./ha spread on nontermite plots just before the experiment began. Dieldrin is a persistent, nonphytotoxic and nonsystemic insecticide of high contact and stomach activity to most insects (Charles 1979). Dieldrin was used after Dursban EC (at the rate of 400 g a.i./ha) had failed to keep termites away from nontermite plots. The main plots were 50 x 50 m and were 50 m apart. The subplots were two groups of four plots, each 15 x 8 m and 8 m apart. The two groups of subplots in each main plot were 15 m apart. Mulch treatments were applied randomly in these plots. These were straw of Pennisetum pedicellatum applied at 3 tons/ha, woody material of Pterocarpus lucens applied at 6 tons/ha, and composite (straw and woody material) treatments applied at 4 tons/ha. In addition there was a control, bare plot with no mulch. Different rates of mulch were used to achieve the same degree of cover of organic mulch on the subplots. Data were collected from August 1993 until November 1995. Termite activity was estimated 6 months after the application of the mulch by counting systematically the number of termite-made voids on four places of 1 m2 selected randomly on each subplot. This gives a good indication of the importance of termite activity. Every year the presence and absence of termites are observed on the plots. Vegetation (herb) parameters were recorded on the plots during three rainy seasons (1993, 1994, 1995). The point quadrat method (Daget & Poisonnet 1971) was used to assess the probability of species occurrence, total cover, and the dynamics of species composition. The dry biomass of plants was assessed at maximum standing crop (September) by the integral cut method, cutting all living plants found in a 1-m2 quadrat. Forty samples were taken from each main plot (five samples per plot), air dried, and weighed. Data of plant dry biomass were used to calculate rainfall use efficiency (RUE) and infiltrated water use efficiency (IUE) for each main plot and each subplot. Rainfall use efficiency is the quantity (expressed in kg) of above-ground phytomass produced per 1 mm of rainfall, and infiltrated water use efficiency is the quantity of above-ground phytomass produced per 1 mm of infiltrated water. Infiltration was estimated as the difference between rainfall and runoff, assuming that surface water storage is part of the infiltration. RUE and IUE serve in dry lands as excellent indicators of soil, and hence of ecosystem productivity (Aronson et al. 1993). In 1995, the woody component of the vegetation was surveyed by systematically counting and identifying all woody species on the plots to establish their density.

Ecological Outcomes Achieved

Eliminate existing threats to the ecosystem:
The rehabilitation of crusted bare soil in the presence of termites + mulch occurred within a short time span. Significant improvement in infiltration was found within 6 months (Mando et al. 1996). Even though our experiment was set up late in 1993 with regard to the onset of the rainy season (June), the vegetation responded to the treatments within the first year, and the response was amplified in 1994 and 1995. In all 3 years the plots with termite activity had more vegetation cover than insecticide-treated plots, but the differences were statistically significant during 1994 and 1995 and not in 1993. Mulch affected the vegetation cover in all the years, but the three mulch types (woody, straw, and composite) were always a homogeneous group that differed from bare plots, except for the first year when woody plots did not differ from bare plots. During the 1994 and 1995 rainy seasons, the data revealed a significant interaction of mulch and termite on plant cover (P = 0.02 in 1995 and 0.05 in 1995). In 1994, termite + straw plots had more plant cover than other mulch + termite plots. At the same time, nontermite plots did not differ from each other and from bare plots. By 1995, nontermite plots and bare plots were still a relatively homogeneous group compared to the termite + mulch plots. Between 1993 and 1995, plant cover increased on all plots except on bare plots where no vegetation developed. The mulch types did not differ much in their effect, but this homogeneous group differed from bare plots, which remained bare during the three years. Termite and mulch interaction did not have a statistically significant effect on biomass in 1994 (P = 0.13). Plant biomass decreased in the order termite + straw, termite + composite, termite + woody material, nontermite composite, nontermite + straw, nontermite + woody, and bare. In 1995, termite and mulch interaction was significant (P = 0.03). Termite mulched plots formed a homogeneous group that differed from the homogeneous nontermite mulched plots and the bare plots group. The order remained the same as in 1994 in nontermite plots, but in termite plots composite and woody plots performed better than straw. Mean values of biomass data showed that plant biomass had increased in 1995 over 1994 on termite + composite and termite + wood plots but had decreased on termite + straw and on all nontermite plots. Rainfall use efficiency was significantly affected by termites in 1994 and 1995, with termite plots having the greatest rainfall use efficiency (p < 0.01). Mean values of RUE increased from 1994 to 1995. Infiltrated water use efficiency was not affected by termites in 1994 but was affected in the subsequent year, with termite plots having the greatest values. In both cases the effect verged on significance at the 5% level. The mulch effect resulted in statistically significant differences between mulch and bare plots in 1994, but no such significance was observed in 1995. In both cases, composite, straw, and woody treatments formed a homogeneous group. There were more plant species on termite plots than on nontermite plots in 1994. All the vascular plants were therophytes. Diversity indexes and plant occurrence probability indicated the extreme dominance of two species in all plots: Aristida adscensionis and Shoenefeldia gracillis. They accounted for over two-thirds of total plant cover in 1994. Three years after the establishment of the experiment, the following woody species were found on termite plots: Piliostigma reticulatum, Pterocarpus lucens, Guierra senegalensis, Acacia senegalensis, Acacia seyal, Terminalia avicinoides, and Zizuphus mauritiana. They are all well adapted to the Sahelian Sudanian conditions (Grouzis 1991), but they succeeded in establishing only on termite plots, without any significant difference in density between the mulch treatments. No tree managed to survive on nontermite mulched plots or bare ground plots.

Factors limiting recovery of the ecosystem:
Vegetation, measured as biomass, coverage, and number of species, performed better in termite straw plots than in termite composite and termite woody material plots during the first 2 years of the experiment, but the vegetation biomass in the latter two treatments increased over the years and in 1995 surpassed the vegetation biomass on the termite straw. This is probably because straw decomposed at a faster rate than woody material, likely due to the higher lignin content of the latter (Berendse et al. 1987). The protective effect of mulch and its biological effect on soil characteristics decline as the mulch decomposes. Tian et al. (1993) have established that termites prefer mulch that decomposes slowly and thus can retain its protective effects longer.

Socio-Economic & Community Outcomes Achieved

Economic vitality and local livelihoods:
Farmers in Burkina Faso and in other areas of West Africa are extensively making use of termite-mediated processes to enhance soil restoration and agricultural production in their farming systems (e.g., the zai/tassa system, where organic material is put into small holes in which termites enhance decomposition and increase water infiltration; see Roose et al., 1992 and Mando et al., 2000).

Key Lessons Learned

Soil crust (especially a structural crust) is a severe constraint to plant development. Valentin (1995) has indicated its effects on seed emergence and infiltration rate. The zero production on bare plots and the very low production in the nontermite mulched plots indicated that neither removing human or animal pressure from already crusted soil nor protecting it against the impact of rain drops or increasing sediment trapped by the many tiny physical barriers due to mulch can rehabilitate the productivity of structurally crusted soil in a short period. The diverse physical effects of mulch alone appear to be ineffective in structurally-crusted conditions, because nontermite mulched plots do not perform better than bare soil.

But placing mulch on crusted soil in the Sahel results in the colonization of this soil by termites. These termites open up numerous voids on the sealed surface of the soil. These voids have a great hydrological significance because they alleviate the infiltration constraint caused by soil crust. In fact, Chase and Boudouresque (1987) stated that a single termite channel of 0.8 cm diameter can sustain a water flow rate of 500-700 ml/minute for 30 minutes. The increase in water infiltration due to termite-made voids leads to an increase in soil water storage (Mando 1997). Furthermore, termite-made voids improve soil physical characteristics such as bulk density, resistance to penetration, and porosity (Mando 1997) and therefore create conditions necessary and sufficient for both woody vegetation and herbs to reestablish.

Termite effects on IUE verged on significance at the 5% level in 1994 and 1995 (p = 0.07 in 1994 and p = 0.04), while termite effects on RUE were highly significant in both years. This indicates that the gap between the performance of termite and nontermite plots is greatly reduced once the infiltration constraint has been removed, suggesting that their effect on soil infiltration is one of the most important mechanisms promoting the revegetation of crusted soil. The preponderant role of infiltration improvement in vegetation rehabilitation has previously been pointed out by Hien (1995), but the important fact here is that the improvement in infiltration brought about by termites is enough to trigger vegetation development resulting in high biomass and 80-200% plant cover within 3 years.

The present study has shown not only how locally available organic resources (i.e. straw, woody materials, manure, etc.) can be used to stimulate regenerative processes, but also how soil-structure degradation results from eradicating native soil fauna (termites in this case). By constructing and opening voids near the soil surface (top 10 cm), and thus facilitating the infiltration of water, termites are a critical component of the ecosystem. Efforts to avoid future land degradation and rehabilitate already-degraded lands should begin with the continual application of organic resources in order to support and maintain termite populations and thereby promote the regeneration of soil and the reestablishment of vegetation.

Long-Term Management

It would be useful to investigate the effect of mulch quality and soil management on the termite population. The efficient mulch rate for revegetating crusted soil and sustaining termite populations, as well as the socioeconomic aspects of using termite-mediated processes in land management should also be explored further.

Other Resources

Mando, A., L. Brussaard, and L. Stroosnijder. 1999. Termite- and mulch-mediated Rehabilitation of vegetation on crusted soil in West Africa. Restoration Ecology 7:33-41.

Primary Contact

Organizational Contact