Shistocerca gregaria, Dociostaurus maroccanus and Locusta migratoria remain the most fearsome orthopteran pests in Algeria. While multiple efforts are engaged in the forecasting and control of S. gregaria, Moroccan and migratory locusts drawless attention. Surveyand antilocust control teams have indicated infestations of D. maroccanus in the western and eastern highlands and of L. migratoria in the Central Sahara. Field investigations reported here show that the initial swarming of both locusts may occur as a consequence of human agricultural practices that make them more successful in their new habitats: increasing locust densities are correlated with human-induced changes in vegetation structure. In this paper anthropogenic impact on locust population density is analyzed using various data methods.
Besides the desert locust, two other major acridid pests, Dociostaurus maroccanus (Thunberg) and Locusta migratoria (L.), are liable to inflict important damage on crops, pastures and forestry resources in Algeria. The Moroccan locust ranges from the Madeira and Canary islands in the west, through the Mediterranean zone (Spain: del Canizo 1950, Cyprus: Dempster 1957) to Afghanistan and south Kazakhstan in the east (Latchininsky & Launois-Luong 1992). In Algeria, swarms of this locust set off from seven gregarization areas, distributed in the highlands in the northern part of the country (Pasquier 1937). The newest outbreak of this acridid started in 1999 and continues in the western zones, where more than 21,000 ha were treated (Anonymous 2000).
As for the migratory locust, Chopard (1945) indicated this species from both the coastal and oasis zones, and also mentioned that Saharan biotic and abiotic factors are unsuitable for L. migratoria to change totally into the economically important gregarious phase from the solitarious one. However, human activity during the last 20 y in the Central Sahara, especially the introduction of irrigated crops under pivoting sprayers, has favored the multiplication of this locust.
In this paper, we evaluate the effect of human alteration of habitats on the swarming behavior of both locust species in two environmental types, geographically distant and ecologically different: the Algerian Sahara and the semiarid zone.
Materials and Methods
Studies of D. maroccanus' populations were carried out in the semiarid highlands between lat 36°N, long 6°W and lat 36°N, long 0 ° W (Fig. 1). Elevation of this breeding area ranges from 1000 m in the west to more than 1400 m centrally. Climatically, the area is classified as a semiarid zone with two distinct winters (Emberger 1971): a severe winter in the eastern highlands and centrally, which is significantly colder than the temperate one in the western highlands. Annual precipitation is from 200 to 600 mm. Alongside some plants resistant to cold, such as Pistachia terebrunthus, Ononis atlantica and Genista myriantha, is a diversified and close-cropped vegetation which develops on compact and clayeysoils after a winter or a spring rainfall. This environment type is characterized by short and overgrazed pastures dominated by Stipa tenacissima, Artemisia herba alba and Ampelodesma mauritanica in isolated communities.
From 1991 to 1992, we investigated two sites situated in the cold semiarid area, in Aïm Boucif region, southeast of Medea (Fig. 1). The first site (El Benia, S1) is an ancient and very open fallow overgrazed by sheep. The vegetation is dominated by herbaceous plants such as Poa bulbosa and Artemisia herba alba. The second site (Aïn Sefra, S2) is also an ancient fallow, where the height and the density of the vegetation is more diverse, with a dominance of Dactylis glomerata, Genista tricuspidata and essentially Asteraceae species. This second site was more moderately disturbed: overgrazing was not so intense and sometimes absent. During 1999–2001 additional investigations were accomplished in many eastern and western regions of the country (Fig. 1) situated in the permanent breeding areas of the Moroccan locust. Most of these zones were remote locations where there was no human disturbance and no grazing by sheep. The plant cover was very closely cropped, punctuated by large bands of bare soil.
The migratory locust was studied in the Central Sahara desert between lat 27°05′N, long 0°08′W and lat 28 ° 15′N, 0° 10′W, at altitudes between 150 m and 300 m (Fig. 2). The climate is typically Saharan, with annual precipitation <100 mm (from the Ghardaïa region to the extreme south of Algeria, Fig.1) (Emberger 1971, Ozenda 1977). Temperatures are elevated from June (33.5° C) till September (31.0°C) [(max.+ min.)/2]. The natural vegetation is rare or absent and ephemeral, but there are perennial plants: Acacia radiana, Tamarix, Calligonum and Rhus tripartitum. In Saharan environments, the presence of L. migratoria is limited to oases; this led us to expect its establishment in the irrigated perimeters created by human agriculture. L. migratoria populations were studied from 1995 to 1997 in 9 sites (Fig. 2, Table 1), distributed along a transect of 150 km through the Adrar region of the Central Sahara, including areas of cereal fields irrigated by pivoting sprayers.
Study of anthropogenic influence on the activity of both locusts.
Human pressure on the swarming levels of the two locust species was assessed by examining the floristic composition of locust habitats, the human activity level in these habitats and by making locust density measurements, the latter taken during the period of maximum activity of the insects: June for D. maroccanus and April for L. migratoria.
Locust sampling and density measurements.
Sampling of D. maroccanus specimens was conducted between 0800 and 1000 in a 1-ha randomly selected area divided into 10 quadrats of 3 X 3 m (Voisin 1986). Sampling of L. migratoria was carried out late in the afternoon along a 100-m transect. The L. migratoria hoppers and adults which flew away or remained in the vegetation, were counted in a 1-m wide band. Population density was calculated following Duranton et al. (1982): larval density ha−1, = (number of observed larvae / number of repetitions) × 10,000; adult density ha−1 = (number of observed adults / sampled surface) × 10,000.
Floristic sampling, global vegetation cover and cover rate of weeds.
At each study site, the floristic composition was determined by visual survey of a randomly placed quadrat of 500 m2. Plant species were censused within this quadrat, with reference to the flora proposed by Quezel & Santa (1962). (Weeds are all non-crop plants: adventitous, including grasses.) A biocenometer of 50 cm per side was thrown ten times in a 1-ha randomly selected area to estimate the different plant cover rates. For each plant, the cover was estimated by calculating the surface area occupied by the orthogonal projection of the aerial part of the considered plant/weed species, across the global area surface, according to Duranton et al. (1982). The total cover rate is the sum of the different rates of each weed species in the study sites. The vegetation structure and physiognomy were studied in April in the southern Locusta sites and in May in the (northern) Dociostaurus sites.
The relationship between anthropogenic impact and locust population level was estimated in the following three ways:
1) By determination of the influence of overgrazing on the density changes in the Moroccan locust.
Crazing impact at the two sites (S1, S2) was assessed during each of our trips to the locales, as the frequency of passage of flocks of sheep and goats. Our observations of the Maroccan locust were obtained between March and September, indications of adults being numerous in the month of June. We undertook two trips per week to each station between 10:00 and 15:00 h.
We calculated Morisita's index (Southwood 1978), as a measure of dispersion of individuals inside the plots at each study site, according to the following formula:
ni = number of the individuals in the i (th) quadrat.
x = sum of all the individuals inside the quadrat.
Is = Morisita's index.
The distribution of the individuals is regular if Is<1, random if Is = 1 and dumped if Is >1.
2) By establishment of a correlation matrix between the L. migratoria population densities and the weed cover rates.
3) By using the Mc Nemartest (χ2) to make 36 cross-comparisons of the vegetation cover and the densities between the two sites. The Mc Nemartest is a significance test for two dependent samples, used when two groups being compared are paired (Cohen 1988).
The vegetation in the two sites (S1 and S2) consisted of 62 and 64 plant species respectively, belonging to 62 families overall. The commonest species are Poaceae (11 to 14 species) occupying 30 to 40% of the ground cover and represented by Poa bulbosa and Dactylis glomerata; also common are the Asteraceae, Fabaceae and Lamiaceae (Doumandji-Mitiche et al. 1993, Benfekih & Doumandji-Mitiche 1996). The floristic richness at site 2 is greater. We found in particular 6 Asteraceae species of which Sonchus oleraceus, Crepis vesicaria and Scolymus hispanicus were very abundant. Twenty six other species of 19 different families form continuous plant groups of approximately 15 to 20 cm height. By themselves these plants occupy 30% of the vegetation cover at site 2. The eastern and western Moroccan locust sites are situated in degraded areas near croplands. They are composed of a similar mosaic vegetation, dominated also by Poa bulbosa.
Table 2 shows the adult densities of the Moroccan locust in June 1992 in the study sites southeast of Medea. At the El Bénia site (S1), adult numbers per hectare were 3 times greater than those at Aïm Sefra. Even though the floristic composition of the two sites is very similar, site 1 presents a dried and overgrazed vegetation cover as a result of the frequent passage of sheep. On each of our trips the passings of flocks of sheep were more frequent at S1 (El Bénia) than at S2 (Aïn sefta): of 8 trips in June, 7 flocks at S1 but only 4 at S2. The number of grazing animals was estimated as about 70 at S1 and several dozen at S2.
The different plant groups at site 1 are often interrupted by rocks and the proportion of bare soil is 45%. Site 2 is more steeply sloped, and so less accessible to sheep than site 1; it shows a denser, more or less continuous, vegetation cover of 74.5% (Table 2). Plants are taller at site 2 and the proportion of bare soil is only 25%. The Morisita index indicates a clumped locust distribution at site 1, where gravid females gather in bare spots to lay their eggs. Adult insect density and plant cover differed significantly (p = 0.05) between the two sites (&chi2= insect densities = 794.34; (χ2) plant cover = 36.83). On the other hand, table 3 data indicate that D. maroccanus has proliferated during the two last campaigns (1999 and 2000) in the eastern and western highlands of Algeria, particularly in Sidi Bel Abbes and Saïda regions (Fig. 1). Swarming levels of the Moroccan locust exceeded 100 young hoppers m−2 and 15 to 25 adults M-2 in more than 20,000 ha at Sidi Bel Abbes and more than 15,000 ha at Saïda were infested by the Moroccan locust between April and June 2000 (Anonymous 2000). Most of these zones are degraded terrains, forest edges and barren desert or ungrazed fallow. In the croplands zone, the locusts inflicted severe damage, notably to barley and soft wheat.
The 9 Locusta sites are all irrigated areas bordering on monospecific introduced cereal crops. The annual flora consists of various plant families: the most represented are the Fabaceae, Asteraceae, Cucurbitaceae, Liliaceae, Cruciferaceae, Ombelliferaceae, Plantaginaceae and Solanaceae. Others, Asclepiadaceae (Pergularia tomentosa), Cyperaceae (Cyperus rotundus), Polygonaceae (Silene setacea, Emex spinosa), Caryophyllaceae (Vacaria pyramidata), Euphorbiaceae (Euphorbia dracunculoïdes), Boraginaceae (Lithospermum appulum) and Plombaginaceae (Limonium thouini ssp blondueli) are less important. Sites 1 and 3 are crops under palms: essentially Triticum durum, Allium cepae, Daucus carota, Lactuca sativa and Vicia fabae. Site 2 is an ancient oasis. Sites 4 to 9 are cereals irrigated by pivoting sprayers. Sites 5, 6 and 7 are the most grassy, with a weed proportion varying from 16% to 26%: 10 to 20% are adventitious grasses. The oasis S2 is a natural habitat with only 6% weeds (Table 4).
In the very different conditions of the arid zone and the Central Sahara desert, the presence of L. migratoria in the 9 study sites appeared to be important in those areas where there is weed cover, especially when one of the adventitious grasses was abundant (Table 4, Figs 3,4). Thus, 78% of the sites with the proportion of weeds <10% had L. migratoria densities between 125 and 3,000 individuals ha−1 (Fig. 5), whereas more than 6,000 to approximately 15,000 specimens colonized sprayer-irrigated areas, with 25% of cover grasses. When only 2 to 4 wild-grass species formed the grass stand and a dominant one had a cover percentage varying from 47 to 67%, total numbers of L. migratoria became high (Table 4, Figs 3,5). The insects were very strongly attracted by Avena sterilis and Sorghum vulgare in particular, even when the proportion covered by these plants was very low—not more than 4% (Fig. 4). Correlation was highly significant between the two types of vegetation cover and the locust population density (0.68 < r < 0.89) as well as for the locust density and Avena sterilis/Sorghum vulgare cover (0.68 < r<0.99). For the 36 cross comparisons between the two sites, the Mc Nemar test reveals significant data at the threshold of α = 5%, (df= 7). The spatial distribution was clumped in all sites except the oasis (Site 2, Table 4).
Discussion and conclusions
According to our results, it appears that both locust species can be negatively or positively affected by anthropogenic disturbance. In the Algerian semiarid area, the swarming risk of the Moroccan locust becomes possible in two situations. First, swarms can develop progressively in regions affected by a certain degree of sheep grazing (disturbed habitat, site 1, southeast of Medea). Skaf (1972) similarly reports the close association of sheep and cattle with D. maroccanus in breeding areas in Syria. Second, the swarming migratory form of the species arises through desertion of fallow areas situated at the bottom of mountains and also, since 1994, through abandonment of nearby marginal croplands by agriculturalists. These human actions lead to the installation of solitarious Moroccan locusts in the abandonedzones, (undisturbed habitat, western and eastern sites). Upsurges which occured in 1999 mainly in the western zones of the country, had thus originally been initiated 3 or 4 y earlier. The absence of phytosanitary measures in the adjacent areas, and the important spring precipitation that appeared in 1999 after several seasons of low rainfall, generated a striking population increase in the Moroccan locust in, notably, the Sidi Bel Abbes region.
Morphometrical measurements based upon many individuals, show characteristic features of the gregarious phase: respectively E/F3= 1.70 and F3/P= 2.59, (E= first pair of wings, F3= posterior femur, P= pronotum). Human impacts were also responsible for the upsurge of D. maroccanus during military conflicts in Tajikistan, at the beginning of 1990 (Latchininsky 1998). The increase of fallow area and drought in the same period in Hungary contributed as well to produce an outbreak necessitating chemical treatments, as mentioned by Nagy (1995).
Agricultural practices in the Central Sahara are another part of the anthropogenic impact on the dynamics of migratory locust populations. In the arid desert zone, L. migratoria has a tendency to concentrate in the Touat region (Fig. 2) where the cropped area has dramatically increased by more than 60% from 1984 to 1989 (Anonymous 1990). This region is a ‘corridor’ of flat land, characterized by rough-texture soils, more or less high in salt content. In this area, humans cultivated cereals under irrigation by pivoting sprayers, and one can find nearby truck and other traditional crops under palms. Most of the irrigated fields are concentrated in the ‘Azzi’ plateau, 70 km south of the Touat region (Fig. 2).
Plant protection specialists found localized but quite dense populations of L. migratoria on irrigated cereals in both the summer and winter of 1994. These populations were not morphologically gregarious in form and color. According to our observations from 1995 to 1997, the activity of the migratory locust was strictly correlated, as expected, with the level of weeds in proportion to crops. Population density went up, when almost 20% of the vegetation cover was represented by adventitious grasses. But the population of L. migratoria only continued to rise beyond this threshold of 20% if Avena sterilis and Sorghum vulgare were present. Damage to plants of Sorghum was noticed at field edges, dose to the cereals irrigated by pivot sprayer systems; and as well damage was observed in the region south of Ghardaïa, in the oazian zone, 600 km north of the Central Sahara, again in the area of pivoting sprayers.
Taking into account the relation between Zonocerus variegatus populations and an imported weed, Chromolaena odorata (Boppré & Fisher 1994), and considering that locusts are attracted to grasses because these plants do not have deterrent chemicals which repel acridids (Bernays & Chapman 1994), it seems relevant to search for the significance of the relation between Sorghum and Locusta under our climatic and especially anthropogenic conditions. This phenomenon, the interaction of agricultural practice and migratory locust swarms, is unique to the Sahara desert. A similar impact was observed with the expansion of pivot irrigation in Lybia in the 1980s. However, up to now, no research has addressed the problem of L. migratoria in the Algerian Sahara.
As a result of Launois' observations in the Lybian desert, the pest species seems to be L. migratoria migratorioïdes (Reiche & Fairmaire 1850), (Duranton, pers. comm). This locust represents a single nondiapausing and polymorphic subspecies from the tropics and the southern hemisphere (Farrow and Colless 1980). Locusta migratoria capito (Saussure 1884) is another migratory locust subspecies, whose outbreaks in Madagascar (Lecoq 1975) depend on a monthly rainfall of 50 to 100 mm. Heavy swarms of L. migratoria were also linked to the extent of rainfall in southwestern Chad during 1988–1989 and early 1997 as mentioned by Chara (1989) and Balança et al. (1999). Studes by Verdier (1972) on the North African Locusta indicate that it is Locusta migratoria cinerascens Fabricius. The bibliography also reveals the presence of this species in the oasis as well as the scrubs and the humid zones (Chara pers. comm). Moreover, it is useful to emphasize that the Locusta of the Adrar region, when reared under laboratory conditions (T= 30°C, photoperiod= 12/12 hours), reproduce continuously, and there is no sign of diapause. Immature Locusta females were also observed in the field 50 km from Adrar (Central Sahara) in December 1995 and January 1996.
Whereas deforestation and drought are factors originally cons idered to be creating suitable habitats (for L. migratoria manilensis in Indonesia (Lecoq et al. 2000)), the colonization of new potentially favorable habitats by L. migratoria (Linnaeus 1758) in the Algerian Central Sahara, appears to result from the expansion of irrigated cereal crops. The use of local nonselected seeds, the absence of crop rotation and the increase in the proportion of weeds under pivoting sprayers are also, up to now, the main factors contributing to the maintenance of the species in this modified habitat, leading to a complete disappearance of this locust from its native habitat situated in the oasis. Determination of the risk of swarming by L. migratoria in the Central Sahara could be possible by following increasing population densities, which we could only understand in the long term after studying the extreme limit of the anthropogenic impacton this locust's population dynamics. Because the locust taxonomy under Algerian ecological and climatic conditions is not precisely known, it is too early to indicate a gregarization density threshold for the species.
In conclusion, we believe that independently of the spring rainfall, humans are creating suitable habitats for the Moroccan locust by overgrazing in the Algerian semiarid area. Consequently, this acridid remains a pest in this zone as was pointed out by Latchininsky (1998). As for the migratory locust, it can be a serious pest in the irrigated modified habitats of the central desert.
The authors wish to thank the personnel of the National Plant Protection Institute at Algiers, and the authorities of Plant Protection Inspection at Adrar for their collaboration. Ourthanks go also to Dr. M. Biche and Dr. A. Eddoud at the National Agronomic Institute at Algiers (El Harrach) for helping with suggestions and statistics. This study is partly supported by a grant from the Superior Education Ministry (EXIPV/1498/97) as well as the Delegate Ministry for University Scientific research (AU/39903). Our special thanks go also to Dr A. Latchininsky (University of Laramie, USA) for his critical evaluation of the manuscript and Dr. C.K. Morris (University of Toronto, Canada) for helping with the english.
Characteristics of study sites of L. migratoria in the Touat region (Central Sahara).
Observations of the Moroccan locust in June 1992 in study sites southeast of Medea.
Observations of the Moroccan locust in 1999–2000 in western and eastern regions of Algeria.
Observations of the migratory locust in April 1996 in irrigated study fields of Adrar region.