Cyclones have been suggested to play a major role in the evolutionary history of the flora and fauna of Madagascar (Wright Reference WRIGHT1999). In spite of the proposed significance of cyclonic activity, very little is known about the specific effects of cyclones on Malagasy ecosystems (Birkinshaw & Randrianjanahary Reference BIRKINSHAW and RANDRIANJANAHARY2007). Understanding the effects of these severe natural disturbances requires comparing data collected both before and after the event. Because cyclones are stochastic, researching the influence of cyclones on Madagascar's biodiversity is difficult and opportunities to study their impact on forests are rare. Nevertheless, studies of the impact of cyclones on Malagasy biota are imperative because models of global climate patterns predict that large-scale tropical disturbances will increase in frequency and intensity in the future (Metcalfe et al. Reference METCALFE, BRADFORD and FORD2008).

The high wind velocity of cyclones and hurricanes can cause substantial damage to tropical forests (Everham & Brokaw Reference EVERHAM and BROKAW1996). Trees are often snapped or uprooted and can experience complete defoliation (Pavelka et al. Reference PAVELKA, BRUSSELERS, NOWAK and BEHIE2003). This damage can have detrimental effects on the wildlife inhabiting the forest by affecting the food supply, nest sites and substrates for locomotion. Because food supply is frequently reduced following the cyclone (O'Brien et al. Reference O'BRIEN, KINNAIRD, NURCAHYO, PRASETYNINGRUM and IQBAL2003, Pavelka & Behie Reference PAVELKA and BEHIE2005), animals often shift their diets to focus on less-preferred food items (Berenstain Reference BERENSTAIN1986), narrow their diets (Behie & Pavelka Reference BEHIE and PAVELKA2005, Tsuji & Takatsuki Reference TSUJI and TAKATSUKI2008) and/or increase their home ranges (Wilson et al. Reference WILSON, GOOSE and WILSON2008). Indeed, many animal populations experience greater mortality as a result of the indirect effects on habitats rather than as a direct consequence of the cyclone or hurricane (Lugo Reference LUGO2008, Wunderle et al. Reference WUNDERLE, LODGE and WAIDE1992). For instance, limited immediate effects after a severe storm were observed at Berenty Reserve in Madagascar: the observation of only a single stunned lemur (Propithecus verreauxi, Rasamimanana et al. Reference RASAMIMANANA, RATOVONIRINA, JOLLY and PRIDE2000).

On 21 January 2009, Cyclone Fanele made landfall on the western coast of Madagascar with sustained winds of 185 km h⁻¹ and gusts up to 260 km h⁻¹, according to the Réunion Météo-France. The dry deciduous forest of the Kirindy Mitea National Park (KMNP) sustained substantial damage as a consequence of this cyclone (Lewis & Bannar-Martin unpubl. data). Ninety-six per cent of the forest trees were found damaged and 9% of the trees were dead approximately 7 mo after the cyclone. Trees with the largest girths experienced the greatest damage and mortality. The understorey and the emergent trees had especially high mortality rates.

Verreaux's sifaka (Propithecus verreauxi, Grandidier 1867) living in the KMNP has been studied since 2006, providing a rare opportunity to study the impact of a major disturbance on sifaka health. The lemur food supply was severely disrupted and altered after the cyclone: 86% of actual food trees were damaged and actual food trees experienced 7% mortality (Lewis & Bannar-Martin unpubl. data). In particular, 25–50% of the largest food trees (e.g. species of Adansonia, Commiphora, Cordyla) experienced severe damage, such as snapped trunks and/or major delimbing. Frequently consumed foods, such as Dalbergia and Baudouinia species, experienced similarly large amounts of severe damage. Therefore, we predicted that the lemurs had experienced a reduction in (1) body mass, (2) body fat and (3) offspring production after the cyclone.

Verreaux's sifaka is a folivorous arboreal lemur (Lewis & Kappeler Reference LEWIS and KAPPELER2005) that lives in the dry forests of western Madagascar. It has a body mass of 2.5–4.0 kg (Lewis & Kappeler Reference LEWIS and KAPPELER2005, Richard et al. Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2000). Verreaux's sifaka lives in small, cohesive, mixed-sex groups of 2–13 individuals (Richard et al. Reference RICHARD, RAKOTOMANGA and SCHWARTZ.1993) in home ranges of 4–10 ha (Lewis Reference LEWIS2005, Nagy Reference NAGY2007, Norscia et al. Reference NORSCIA, CARRAI and BORGOGNINI-TARLI2006, Richard et al. Reference RICHARD, RAKOTOMANGA and SCHWARTZ.1993). Home-range use decreases significantly (Norscia et al. Reference NORSCIA, CARRAI and BORGOGNINI-TARLI2006) and consumption of mature leaves increases dramatically (Lewis & Kappeler Reference LEWIS and KAPPELER2005) during the dry season, causing the lemurs to lose 10–20% of their body mass (Lewis & Kappeler Reference LEWIS and KAPPELER2005). Mean skinfold thickness, an estimate of subcutaneous body fat, decreases from 4.5 mm to 1.1 mm during the dry season (Lewis & Kappeler Reference LEWIS and KAPPELER2005).

Verreaux's sifaka has a limited mating season, January–March (Brockman Reference BROCKMAN1994, Lewis & Kappeler Reference LEWIS and KAPPELER2005). During the dry season when its diet is almost entirely mature leaves, females give birth to one infant, primarily in July–September (Lewis & Kappeler Reference LEWIS and KAPPELER2005, Richard Reference RICHARD, Martin, Walker and Doyle1974). Females then take a ‘leap of faith’ that sufficient resources will be available for the most energetically demanding reproductive phase of late lactation (Lewis & Kappeler Reference LEWIS and KAPPELER2005). Females with larger body masses during the mating season have a higher probability of giving birth and successfully rearing the infant (Lewis & Kappeler Reference LEWIS and KAPPELER2005, Richard et al. Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2000).

The KMNP is one of the largest protected areas in Madagascar with 140 000 ha. The park encompasses the transition between three ecosystems: western dry forest, southern spiny desert and mangroves. This study was conducted within the 1-km² Ankoasifaka Research Station at 20°47.69′S, 44°9.88′E.

Between 2006 and 2010, adults from five social groups were individually marked with nylon collars and radio collars and captured multiple times as a part of a long-term study. Capture methods have been described in detail elsewhere (Lewis Reference LEWIS2009, Lewis & Kappeler Reference LEWIS and KAPPELER2005) and are summarized here. Animals were captured using a blow gun that delivers disposable non-barbed darts with a 9.5-mm needle. Darts were loaded with Telazol™, which is a non-narcotic, non-barbiturate, injectable anaesthetic. Sedated individuals had their body mass measured with a 5-kg spring balance and body condition was measured with skinfold-thickness calipers (Slim Guide, Creative Health Products). Animals usually recovered and returned to their social group within 2–4 h.

Skinfold thickness can be a good proxy for subcutaneous body fat in primates (Hamada et al. Reference HAMADA, HAYAKAWA, SUZUKI, WATANABE and OHKURA2003, Walker et al. Reference WALKER, SCHWARTZ, WILSON and MUSEY1984). While it has yet to be validated with actual measurements of body fat in Verreaux's sifaka, its measurement can provide useful information about the relative change in body condition. Skinfold thickness measurements from the hip (skinfold over the inguinal ligament), abdomen (the centre of the abdomen where the hair swirls), arm (centre of the back of the arm) and back (just below the shoulder blade) were collected by RJL and summed to produce a composite score. These scores have been shown to follow the same seasonal variation as body mass in sifaka (Lewis & Kappeler Reference LEWIS and KAPPELER2005).

Only three adults captured in 2006 were recaptured in 2010. Also, the animals were unable to be captured in 2009 immediately after the cyclone due to the forest damage. Hence, this analysis compares the body mass and skinfold thickness of the six adults that could be recaptured in late June–early July in 2007, 2008 and 2010 as well as three additional adults that were captured in both 2007 and 2010, for a total sample size for body condition of nine adults.

Demographic records from the long-term study of the behavioural ecology of Verreaux's sifaka were used as a proxy for female condition because females in poor condition are not expected to be able to reproduce. Demographic data were collected via monthly censuses and supplemented using the more intensive behavioural observations. Birth records were available for nine females in the five social groups (including the five females in the body condition dataset). While data were available for all five years (2006–2010), the birth rates of four females could only be examined for four years because in 2006 either they were not of reproductive age or they had not yet been included in the study.

Although sifaka weighed more in 2008 than 2010 (Table 1), body mass did not significantly vary across the three study periods (repeated-measures analysis using a linear mixed model: F_2,13.15 = 0.189, P = 0.830). By contrast, skinfold thickness did significantly vary across the three study periods (repeated-measures analysis using a linear mixed model: F_2,13.05 = 23.5, P < 0.001). A comparison of years with 2010 revealed that this result was driven by the increased skinfold thickness in 2008 rather than reduced skinfold thickness in 2010 (2007 vs. 2010: t_12.73 = 0.606, P = 0.555; 2008 vs. 2010: t_12.73 = 6.42, P < 0.001).

Table 1.

Yearly variation in Verreaux's sifaka body mass, skinfold thickness and the per cent of females with infants during the birth season in the Kirindy Mitea National Park of western Madagascar. Cyclone Fanele passed over the dry deciduous forest in January 2009. Measurements reported below were collected in late June–early July of each year. Means are presented with standard deviations. Birth results are based on the same nine females from 2007–2010.

A mean (SD) of 61% ± 14% of females gave birth in any given year (2006: 3/5 females, 2007: 4/9 females, 2008: 7/9 females, 2009: 6/9 females, 2010: 5/9 females). Females exhibited a mean (SD) birth rate of 0.59 ± 0.22 infants y⁻¹. While the percentage of females giving birth in 2010 was lower than 2008, it was not the lowest observed (Table 1).

Animals respond differently to natural disasters (Waide Reference WAIDE1991). For example, the grey-headed lemur (Eulemur cinereiceps) demonstrated resilience in response to a cyclone in Madagascar (Johnson et al. in press), whereas the brown lemur (Eulemur fulvus) in the Comoros experienced a population decline after a cyclone (Tarnaud & Simmen Reference TARNAUD and SIMMEN2002). Reproduction can also be altered as a result of major disturbances. Vasey & Borgerson (Reference VASEY and BORGERSON2009) found that the red variegated lemur (Varecia rubra) in Madagascar experienced ‘post-cyclone reproductive failure’.

One of the best-studied primate responses to a major natural disaster was conducted in a population of the howling monkey (Alouatta pigra) after Hurricane Iris (Behie & Pavelka Reference BEHIE and PAVELKA2005, Pavelka & Behie Reference PAVELKA and BEHIE2005, Pavelka & Chapman Reference PAVELKA, CHAPMAN, Estrada, Garber, PAVELKA and LUECKE2006, Pavelka et al. Reference PAVELKA, BRUSSELERS, NOWAK and BEHIE2003, Reference PAVELKA, MCGOOGAN and STEFFENS2007). After an initial reduction of 42%, the howler population declined a further 12% 3.5 y after the hurricane due to drastic reductions in food supply. Unlike the sifaka in the KMNP, the howling monkey females did not reproduce for at least 1 y after the hurricane (Pavelka & Chapman Reference PAVELKA, CHAPMAN, Estrada, Garber, PAVELKA and LUECKE2006).

Verreaux's sifaka also experienced significant changes to its food supply and in the overall forest structure following Cyclone Fanele (Lewis & Bannar-Martin unpubl. data). Nevertheless, this lemur population did not experience a drastic post-cyclone reduction in population density (2008 vs. 2010: 42 vs. 37 individuals km⁻²: Lewis unpubl. data). Moreover, sifaka health, as assayed with body mass, skinfold thickness and births, did not decline in the first 2 y after the cyclone. While health may have deteriorated immediately following the cyclone and then rebounded 18 mo later, the birth data 8 mo after the cyclone suggest that female condition was not severely affected by the cyclone. Indeed, our results are comparable in all three years to the multi-year values reported for a much larger sample of Verreaux's sifaka from Beza Mahafaly Special Reserve (23–77%: Richard et al. Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2000, Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2002).

Verreaux's sifaka lives in highly seasonal habitats (Lewis & Kappeler Reference LEWIS and KAPPELER2005, Richard et al. Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2000) and exhibits a seasonal breeding strategy that times lactation, the most costly phase of reproduction, with the period of low and increasing food availability (Lewis & Kappeler Reference LEWIS and KAPPELER2005). Because its body condition deteriorates substantially during the lean season, females are not always able to recover in time to cycle the following mating season (Lewis & Kappeler Reference LEWIS and KAPPELER2005, Richard et al. Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2000, Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2002). It also exhibits a strong link between environmental changes and body mass (Richard et al. Reference RICHARD, DEWAR, SCHWARTZ and RATSIRARSON2000). Thus, our observed lack of negative responses to the post-cyclone environmental change is unexpected. Moreover, our finding that sifaka in 2008 (before the cyclone) exhibited the highest body mass, skinfold thickness and number of births suggests that this year probably had greater food availability. Because we have only limited rainfall data and no phenological data from 2008, we are unfortunately unable to know for certain.

While folivores suffer as a result of extensive defoliation in their habitat immediately following a cyclone (Lugo Reference LUGO2008, Wilson et al. Reference WILSON, GOOSE and WILSON2008), they may experience an increase in food supply as trees direct their energy towards growth and leaf production and away from reproduction (Vandermeer et al. Reference VANDERMEER, BRENNER, GRANZOW and CERDA1998). Thus, our finding that 18 mo after the cyclone the sifaka did not exhibit significant deterioration in body condition may not be surprising. This potential increase in food supply could also explain the lack of reduction in births following the cyclone. Similarly, a study of three species of arboreal and folivorous ringtail possum (Pseudocheiridae) in Australia found no changes in body mass after Cyclone Larry (Wilson et al. Reference WILSON, GOOSE and WILSON2008).

Long-term data are required to assess the full impact of Cyclone Fanele on the Verreaux's sifaka in KMNP. Nonetheless, this analysis of the immediate effects of the cyclone on the sifaka population suggests that it was not severely impacted by the cyclone. Our results are in accord with the Wilson et al. (Reference WILSON, GOOSE and WILSON2008) hypothesis that medium-sized folivores can be resilient to even severe cyclones, and with the Wright (Reference WRIGHT1999) energy conservation hypothesis that lemurs are well-adapted to coping with disasters and conserving energy.