NAT. NAT. HIST. BULL. SIAM Soc. 38 :99 ・115 ， 1990

FIRE ，DRY DIPTEROCARP FOREST ，AND THE CARNIVORE COMMUNITY IN HUAI KHA KHAENG WILDLIFE SANCTUARY ，.

Alan Rabinowi 貯

ABSTRACT

Widespread Widespread annual burning in p紅白 of Tha iland has degraded natural forest

forrnations forrnations and encouraged 白e sp 隠 ad of 血e fire resilient dry dipterocarp forest. Th e influence influence of 白草加d dry dipterocarp forest on 白 e carnivore community w ぉ exarnined

during during a two year study in Huai Kha Khaeng Wildife Sanctu 創 y ，Th ailand. Seasonal fires influenced influenced mortality and movements of carnivore species. All carnivores showed a dispro- portionately portionately low use of dry dipterocarp forest ∞ mpared to other habitat types ， and a1 1 took

問 fuge in evergreen forest during the driest tirnes of the year. Some carnivore species were res 甘icted to evergreen forest alone. Th e evergreen forest was a crucial component of

carnivore carnivore home r組 ges. The dry dipterocarp forest was a resource poor habitat character- ized ized by seasonal water scarcity and a relatively low abundance of prey species and fruiting

U官es. Small carnivores were mo 陀 a釘'ected by fires and by 問 source scarcity in dry diptero-

carp carp forest than we 陀 larger ，wider-ranging carnivores. However the dis 町ibution of the in in the study area was also limited by 也 y dipterocarp fores t. Despite a diverse carnivore

comrnunity comrnunity in the study area ，se ぉ onal 白res and 由ep 陀 sence of dry dipterocarp forest threatens threatens the heal 出 and stability of many carnivore populations. Dry dipterocarp forest should should not be allowed to spread at the ex 戸nse of evergreen forest formations. Total exclusion exclusion of fire is not feasible in Th ailand's remaining forest areas ， but an altemative policy of of prescribed burning combined with fire exclusion in high priority habitats appe 創"S sensible.

INTRODUCTION

Although frre is a naturally occurring phenomenon ，most of the fires 出at have influenced present day tropical forest formations have been anthropogenic (BUDOWSKI ， 1966; BATC 阻 LDER ， 1967). Thr ough a long history of uncontrolled human use ， fire has

become an endemic ecological force in the seasonally dry forests and gr 出 slands of the tropics tropics (BARTLE 廿， 1956; STOπ ， 1988b). Th is is clearly evident over large areas of mainland south -e ast Asia. Widespread annual burning during the long dry seasons has gradually degraded natural forest formations and encouraged the spread of fire-resilient 合y dipterocarp forest formations which today cover large areas of Burma ，Th ailand ，Laos ， Cambodia ，and Vietnam (STOTT ，1984; WHARTON ，1966 ， 1968).

市 Wildlife Conservation In ternational ，New York Zo ological Society ，Br 叩 x，New York 10460

99 1∞ ALAN RABINOWITZ

Despite Despite its destructive force ，fire is rarely considered to be hannful to wildlife. Most vertebrates are capable of escaping hann (KOMAREK ， 1969; S町 GER et al. ， 1989) and do not often die in wildfires or prescribed bums (BENDELL ， 1974; HOWARD et al. ， 1959; L 田 GE. 1968). In areas with seasonal frres ， animals are presumed to be fire-adapted through natura1 natura1 selection (KOMAREK ，1969) ， with avoidance being the most prevalent adaptation (STOπ ， 1988b). Some species of wildlife ，particularly ungulates ，benefit from the protein- rich rich shoots which sprout after a bum (KOMAREK ， 1969; MOE et al.， 1990) and 紅 'e at 位acted to to habitats that fire helps sustain (STOπ ， 1988a; WHARTON ，1966 ， 1968). 百lere is little ioformatioo published on the effects of uncontrolled seasonal fires and fire-associated habitats 00 wildlife communities. In south-east Asia ， despite the apparent apparent benefits of fire and dry dipterocarp forest to large species of ungulates such as gaur gaur (Bos gaurus) and banteog (B. javanicus) ， the dry dipterocarp forest is relatively impoverished impoverished in wildlife such as small terrestrial mammals (W ALKER & RAB 別 bWITZ ， in press; press; W 江田， 1981) ， birds (WILES ， 1980; ROUND ，1988) ，herpetofauna (HEYER ， 1970) and invertebrates invertebrates (WHARTON ， 1966). In the combined 釘 'eas of Th uog Yai 佃 d Huai Kha Khaeog Wildlife Sanctuaries on the westem border of central Th ailand ， only eight verte- brates brates are listed as primarily associated with bamboo and dry dipterocarp forest areas ， while while 325 species are associated with mixed deciduous and evergreen forest formations (NAKHASA' 羽田N & STEWART -C OX ， 1990). In In the dry tropical forests of Hu ai Kha Khaeng Wildlife Sanctuary ，fires are started started by both local people and sanctu 釘 y staff every ye 低 During years of heavy rainfall ， fire fire is primarily confioed to the more open dry dipterocarp forest formations. During drier ye 釘 s，p訂 ticularly when leaf litter and ground cover from previous years has not been bumed ，hotter and more extensive fires penetrate ioto the evergreen and mixed deciduous forest 釘 'eas 佃 d do more damage to 白e dry dipterocarp forest itself (STOπ ， 1986). Fire in in the less fire-resilient mixed deciduous and evergreeo forest ， allows the further spread of of dry dipterocarp formations (Sτoπ ，1986 ， 1988). Research Research on 白e camivore community was carried out in Huai Kh a Khaeng Wildlife Sanctuary Sanctuary between 1987 and 1989 (RABINOW 汀Z，1989 ，1990 ， 1991 ; RAB 町 OWITZ & W ALKER ， 1991). 1 present data on how fire and the dry dipterocarp forest affects some of of these camivores ，and what the cu 町'ent fire regimes mean for the future of the sanctu- ary's ary's wildlife communities. Possible fire management strategies for Th ailand are also discussed. discussed.

STUDY AREA

Huai Kha Khaeng Wildlife Sanctuary (2 ，575 km 2) encompasses 出e Huai (= stream or or small river) Kh a Kh aeng and part of the Huai Th ap Salao watersheds. 官le core study area area encompassed 50 km 2 arouod Khao Nang Rum Research Station ， in the eastem por- tion tion of the sancωary ，and included part of two s悦 am systems ，出 e Huai Chang Tai and Huai Ai Yo. Only the Huai Ai Yo is a permanent source of water throughout the year. FIRE ，DRY Dl PTEROCARP FOREST ，AND CARNIVORES 101

~ Dry evergreen forest 医望 Dry dipterocarp fore st

_ Hill evergreen fore st ~ Mi x巴d deciduou s fore st

直ヨ Field station ーく Stream

Fi gure 1. M ap showin g di stribution of habit ats at slU dy sit e， Illu stration repr ese nt s a 10 km x 10 km area .

Most of the study area range s in elev ation from 400-600 m ， but includes part of Khao Khieo mountain at 1350 m . Vegetation Vegetation in the area is a mosaic of four fore st types (Fig . 1) ， described in greater greater detail by BHUMPAKKAPUN et al. ，(1985) ，STOTT (1 984) ，and THITATHAMAI くUL 102 102 ALAN RABINOWITZ

(1985). 百lese include the following: Mixed deciduous forest ， comprising 35% of the study study area and found primarily on moderately sloping and flat 紅 eas ne 紅 streams. 官le ta1 1est 仕'ees are 20 ー30 m high and average ground cover is 70%; Dry deciduous diptero- carp carp forest ， comprising 23% of the study area ， with a mo 陀 open canopy ，and a floral composition composition indicative of an artifici a1 climax maintained by annual fires (ST O' π ， 1986).

The tallest trees 訂 e2 0- 30 m high and average ground cover is 80% in the rainy season. After After dry season fires ， the ground is mostly bare; Dry evergreen forest ， comprising 35% of of the study area ， a relatively dense formation found mostly along the waterways. Ground cover cover averages 47% ，and is dense only in gap 紅白s; Hill evergreen forest ， comprising 7% of of the study area ， is found only above 1000 m around the summit of Khao Khieo Mtn. Temperature Temperature and rainf a1 1 records from 1983-1987 (Table 1) show clear season a1 pattems. pattems. During the 1988 study season ，rainfall was greater than in previous years. Th e fire fire season ，gener a1 ly from December through March-April (STOTT ，1986) ，was shortened to to March through mid-April by unusu a1 ly heavy rains. Consequently ， there was less fire damage in the study area ，and the dry deciduous dipterocarp forest was bumed more severely severely than the mixed deciduous and dry evergr 白 n fores t.

RESULTS

Twenty-one camivore species of five families occur in the study area (RABINOW- I立 &WAL 阻 R， 1991). A total of 27 camivore species are listed for the entire sanctuary (NAKHASATHIEN & STEWART -C OX ， 1990). Felidae Felidae -One resident tiger (Panthera tなris) and four resident (P. par- dus) dus) were tracked for more than a year (RABINOWIτ2 ， 1989). Despite minimum home ranges ranges of these large cats between 11-37 km えtheir core activity areas (where an animal was located >75% of the time) were much more restricted and were based around water- ways in the evergreen forest formations. During the dry season ， the cats clearly avoided forest forest that was buming and were found in dry dipterocarp forest much less frequently than during during other times of the yea r. Four species of small to medium-sized cats also occur in the study area (RABI- NOWITZ & WALKER ，199 1)， but all except the cat (Felis bengalensis) were uncom- mon. Th e (Neofelis nebulosa) and the jungle cat (F. chaus) were both seen seen on a single occasion in mixed deciduous fores t. The Asian golden cat (F. temmin- icki) icki) was seen scavenging a dead banteng carcass in the dry dipterocarp fores t. Five Five leopard cats were captured and four were tracked for periods ranging from one to 13 months (RAB 町 owr 官， 1990). All capωres and all core activity areas were in dry dry evergreen or mixed d配 iduous forest close to a waterway.

One male shifted during ranges the fire season. For unknown reasons ， he moved from an unbumed evergreen forest area into dry dipterocarp forest that was buming extensively. He took refuge in a ravine along a small waterway but was later FIRE ，DRY DIPTEROCARP FOREST ，AND CA RNIYORES 103

Fi gllr e 2. Fil 巴一 infJ ll enc巴d clr y dipl eroc 山 p fores l in Hll ai K ha K haeng Wi ldlif e Sa nclll ary ( pholo by Se llb Na kha salhi en)

Figllr e 3. Dr y eve rgree n for es l around K h ao Nι II1 g Rlllll R 巴sea rch Slali on in トIllai Kha Khaeng Wildlif e San clU ary ( photo by A l an Rabin ow il Z) 104 ALAN R ABINOW ITZ

Figure 4. Dry se ason fires burn ing dry di pterocarp fores t in Huai Kha Khaeng Sanctuary (photo by Sompoacl Srikosamatara).

Figure 5. An advancing groun d fire in dry dipteroca rp forest in Huai Kha Khaeng Sanctuary (photo by Belinda Stewart - Cox). FIRE. DRY DIPTEROCARP FOREST ，AND CARN IVORES 105

F igur e 6. Th e res ult of s巴ve ral fir e seaso ns ea tin g away at th e "ya ngo il " hol es at th e b ase of a lar ge ，h 巴alth y dipt erocarp (p hoto by Su san Wa lke r) 106 ALAN RABINOWITZ

Figure 7. Chief of Huai Kha Khaeng Sanctuary, Mr. Seub Nakhasathien,.examining fire damage and "yang o il " holes at the base of a !age, healthy dipterocarp (photo by Beli nda Stewan-Cox). FIRE, DRY DIPTEROCARP FOREST, AND CARNIVORES 107

Table 1. Monthly temperature and rainfall at Khao Nang Rum Research Station, Huai Kha Khaeng Wildlife Sanctuary, Thailand, from 1983-1987, and for the study period 1988.

Temperature ("C) Rainfall (mm) Month 1983-87 1988 1983-87 1988

January 20.3 20.7 2.6 0.0 February 23.8 25.3 25.2 112.0 March 1 25.7 26.8 45.2 36.8 April 1 27.5 27.8 105.1 104.1 May 26.6 27.1 229.9 294.6 June 27.1 26.7 133.1 93.2 July 25.9 26.7 109.3 330.0 August 25.7 26.0 114.9 227.0 September 25.2 25.4 302.3 603.5 October2 24.2 24.6 335.0 402.0 November 22.2 21.2 91.6 10.7 December 19.2 19.2 13.5 0.0

Average 24.4 24.8 125.7 184.5

1 Fire in study area in 1988 season. 2 Intermittent flooding in lowlands in 1988 season.

found dead inside a stand of burned bamboo. The cause of death was not apparent but his skeletal remains showed the loss of the upper left canine, an injury not present at capture. A female leopard cat, followed for 13 months, used a total area of 6.5 km2 consisting of 32% dry dipterocarp forest. Despite seasonal shifts in her ranging pattern, all core activity areas were in dry evergreen forest with good water sources. All but one of the eat's located resting sites (N=I8) were on the ground with an average ground cover of 90%; all but three resting sites were in mixed deciduous or dry evergreen forest. A third male leopard cat was followed for only four weeks during December and January, the driest time of year (Table 1), before being killed by camp dogs. During this time, the eat's range of 1.5 km 2 contained 47% dry dipterocarp forest with patchy water sources. This leopard cat was found to be catching free-ranging domestic chickens close to camp. Three other radio-collared leopard cats, in areas with better water sources and more evergreen forest close to camp, were never found taking chickens. 108 108 Al.AN RAo INOWITZ

Two cat species ， the fishing cat (F. viverrina) and the marbled cat (F. marmora ω) ， were not recorded 泊 the study area. But 也ey were known to occur in adjacent 紅'e as of evergreen evergreen and deciduous forest with 1釘 ger ，more permanent waterways.

Viven 也tae - Five individuals of four civet species were radio ・coll 釘 'ed 叩 d fol ・ lowed for periods rang 泊g 合'om six to 12 months. Observations were made on two additional additional species (RAB 町 OWITZ ， 1991b). Home r阻，ges va 釘ri おed between 3.1 km ♂2 for the small small Indian civet (Vi 附iv 陀F氾er 庁r，"ic 伊， zj めbet ，幼ha，ω) ， but seasonal activity centers were nearly all based around waterways in 出edry evergr 四 n fores t. Al though the civets had a wide range of fl ∞d preferences ，including 加出animal prey prey and at least 18 species of 企凶出.g b 田 s (RAB 町 OWITZ ，1卯1)，白 e 命Y dipte J1侃紅 p forest forest w 酪 used propo 凶 onately much less 由加 all other forest types. Eighty pe 問 ent of civet civet day beds located were in mixed deciduous forest less 白an 150 m from a waterway. During During the peak of fire season ，March 伽'O ugh mid-April ，civets had slightly lower 出組 average average levels of activity with larger monthly home ranges 組 d movements. 百le extent of 佃 individual's movements during 由民 time depended on the pro 即 rtion of 命y diptero-

C 紅 p forest in his r加 ge. A male small Indian civet whose range consisted of 62% dry dipterocarp forest

showed his largest mon 白ly home range shift when 自民s came 由rough the 紅 'ea in Ma J1C h. His His movements ， which were previously centered in dry dipterocarp forest ，now shifted into into the dry evergreen forest along the Huai Chang Ta i. A male common palm civet (Paradoxurus (Paradoxurus hermaphroditus) ，whose original range consisted of 60% dry dipterocarp forest 姐 d was centered around Huai Chang Tai ， also shifted his movements when fires

came 伽'O ugh. He established a new range consisting of 44% 合 y dipterocarp fo 問 st which

W 儲 b踊 ed ne 釘 the larger ，more permane 白t waterway ， Huai Ai Yo. A masked palm civet (Paguma larvata) with only 3% dry dipterocarp forest within within his range showed larger 白an average daily movements during the fire season ， but did did not shift away from his pre-frre season area. Hi s range ， with 67% dry evergreen forest ， was virtually untouched by the frres.

百le semi-arboreal binturong (Arctictis binturong) 佃 d the arboreal small-toothed palm civet (Arctoga /i dia trivirgata) were sighted only in the dry evergreen forest compo- nent nent of the study area. Th e banded linsang (Prionodon /i nsang) ，know from evergr 明 n areas areas further west in 白e sanc 佃町 (NAKHASA' 叩 I町 &S 百 WART --C OX ，1990) ，was never reported reported in the study area. Urs 嗣oe- 百le two Asian bear species ，出 e Malayan sun bear (Helarctos malay-

anus) 佃 d the Asiatic black bear (Selenarc ωs thibetanus) ，∞C 町 in the study 創'e a， but bo 由 釘'e uncommon (RAB 町 owrrz& WAL 阻 R ，1991). Occasional be 釘 feces were found along trails trails in the evergreen forest during 白e rainy season. Muste 肱tae - Evidence of small-clawed otter (Aonyx cinerea) was found in 由e evergrl 偲 n forest along a seasonal 剖 of butary Huai Chang Tai. No evidence of the smooth- coated coated otter (Lutra perspicilla) w 邸 found ，al 血ough it ∞C町 s in adjacent areas of mixed 日RE ，DRY DIPTEROCARP FOREST ，AND CARNIVORES 109

deciduous deciduous and evergreen forest containing larger ，more permanent waterways (RAB 町 OW- I1Z & WAL 阻 R， 1991).

DISCUSSION

Dry Di pterocarp Forests

Th e dry dipterocarp forests of south-east Asia are distinctive formations ，charac- terized terized by xerophytic dipterocarps which 釘 e clearly adapted to and often ecologic a1 ly dependent dependent on dry season fires. In the absence of fire ， the dry dipterocarp forests would be be confined like "island refuges" to their edaphic cores ，su 汀ounded by more extensive formations formations of dry evergreen and mixed deciduous forest (STO 廿， 1988b; WHARTON ， 1966).

In In Thailand ， dry dipteroc 釘 p forest occupies nearly half of the forest area (STO 甘， 1988a).

ST O' π(1986) cl a. ssified fires in the dry dipterocarp forest into two types: litter bums and ground cover bums. 百lese bums could be either "typical" or "extreme" depend- ing ing on the depth of litter or the abundance of highly flarnmable pygmy bamboo. Th e main fuel fuel for fires is provided by leaves shed during the early p訂 t of the dry season (up to 90% of of the leaf cover by the end of January) and ，later in the season ，by the ground cover of dηgrasses. Although frequent bumings may restrict seedling ，sapling and tree growth in the the dry dipterocarp forest (SUTTIVANITCH ，1989) ，STOTT (1986) concluded that most types of of fire do little direct damage to the dry dipterocarp forest ecosystem. Only extreme ground ground cover bums ，which occur when an area has not been bumed in a long time ， pose a serious threat to 合y dipterocarp forests.

Dry deciduous dip 戸teroca 釘rp forest comprised nea 紅r匂 one- 叶qu 訂 t旬er of the habitat mosaic mosaic of the study 釘are 伺a and was defined prim 釘 ily by tlωopographi にca 剖 1 features such as st 佐re 創 n beds and sharp ridges (伊 F日抱ig. 1). Season a1 fires help maintain this dry dipterocarp forest forest in a sub-climax state. But even when fires do minimal damage to the floral components of the dry dipterocarp forest ， they con 甘ibute to a cycle in which the forests become drier ，more open ，and their understory more flammable (NAKHASATHIEN & STEWART -C OX ， 1990). Fires a1 so promote the intrusion of the dry dipterocarp forest into the the fire-sensitive dηevergreen and mixed deciduous formations. Although dry diptero- carp carp forest is "clearly adapted to the ancient force of fire" (STOTT ，1988b) ， the continued spread spread of this forest type outside of its edaphic core 釘 'eas threatens the integrity of natural forest forest systems.

The Carnivore Community

In In the ex 汀'eme ，season a1 fires can affect individual mortality ，particularly when other other stresses 釘 'e presen t. This may have been the case for the dispersing m a1 e leopard cat 血at was found dead in a newly bumed bamboo patch ，and was missing one of his canines. canines. Fire can also s汀ess healthy individuals when dry dipterocarp forest makes up a large large proportion of an anima l' s home r佃 ge. Since this dry ，open forest formation is 110 110 ALAN RABINOWITZ

bumedmore f詑quently th 加 other forest types ， camivores must be able to move to escape the the fire ， and must cope with an immediate decrease in food availability. Post-fire 釘'apping in in the dry dipterocarp forest showed that the biomass of three major small mammal prey species species dropped 56% compared with pre-fire 回.p ping at the same site (W ALKER & RAB 別 owm ， in press). Over the long term ， smaller camlvores must cope with the impoverishment of the dry dry dipterocarp forest ，namely seasonal water scarcity and a relatively low abundance and biomass biomass of small terres 回al mammalS (WILES ， 1980; W 札阻R & RASINOWIπ ， in press)

and fruiting 田 es (RAs 町 owrπ ， 1991). Wider ranging movements 訂 e expected when critical critical resources are seasonally affected by fire (CH 則 STIAN ， 1977). Th is particularly important important with small ，solitary ，flesh-eating camivores such as leopard cats ，whose dietary needs needs already force these species to use larger areas than other species of similar size (GITILEMAN & HARVEY ， 1982). A 1 釘 ge proportion of dry dipterocarp forest in the home range range of one of the male leopard cats ，may have prompted his chicken depredations during the the dry season. Larger ，wider-r 組 ging camivores are less affected by the immediate consequences of of fire ， but must also deal with seasonal food and water scarcity in the dry dipterocarp forest. forest. Le opards were not uncommon in the study area but were scarce. Species that that are most successful in surviving and adapting to degraded habitats are those which

combine dietary breath and ranging flexibility (BE 阻 NSTA 凹， 1986). Le op 釘 ds generally

hunt hunt smaller prey ， can use secondary growth and bumed 釘 'eas ， and can survive with less water water than the tiger (J m 町 SINGH ， 1983). Tigers prefer riverine forest with shade and

dense dense vegetation ，and are 出 where larger prey species (5 0- 100 kg) are available (SCHALLER ， 1967 1967 : SUNQUIST ， 1981). In the more extensive evergreen and deciduous forests of the Huai Kh a Khaeng river valley ，tigers were at least three times as abundant as they were in in the study area (RAs 町 owm ， unpublished data). In In the study area ， prey within 出巴 tiger's preferred size range is limited primarily 旬 S制加deer (Cervus unicolor) and wild boar 印刷 scrofa). Sambar deer are seasonally abundant ，becoming sparse and patchy in 出e 企 y dipterocarp forest during the dry season (SRIKOSAMATARA ， unpublished data) ， while wild boar are uncommon in the area because 由ey prefer wetter forest and more permanent waterways (LEKAGUL & MCNEELY ， 1977). Barking Barking deer (Muntiacus muntjak) ， the major prey for leopards ， are relatively abundant in 出e 命 y dipterocarp forests and helped con 住ibute to a healthy leopard population. Impo 託加t secondary secondary prey items the for leopard however ， such as monkeys (Macaca sp. & Presbytis sp.) sp.) and hog badgers (Arctonyx collaris) ， are found primarily in mixed deciduous and evergreen evergreen forest (LEKAGUL & McNEELY ， 1977).

Th e forest mosaic of the study 訂 ea supports a relatively diverse camivore community. community. Ope n woodland species ， such as the common palm civet ， are found alongside dense dense forest species such 部出 e binturong. Yet ，some camivore species ， such as tiger ， medium-sized medium-sized cats ，bears ，and the small-clawed otter ，were uncommon; evidence of their presence presence was often res 住icted to 由e evergreen and deciduous forest. Other species such FIRE ，DRY DIPTEROCARP FOREST ，AND CARNIVORES 111

価 the small-t 'O'O thed palm civet and the bintur 'O ng did n'O t appear t 'O use 白e dry dipter 'O carp f 'O rest at al l. Species 白紙 needed extensive ye 釘 '-r 'O und water s'O urces ， such as the fishing cat cat and the smooth-c 'O ated 'O tter ， were absent from the study area. All All the carniv 'O res studied sh 'O wed a disprop 'O rti 'O nately l'O w use 'O f dry dipter 'O carp f 'O rest c'O mp 釘 ed with 'O ther f 'O rest types within their h'O me range ， and all f'O und refuge in wetter wetter ar 官邸'O f everg 陀 en f 'O rest during the driest times 'O f the ye 抵 Clearly ，the 合y dipterocarp dipterocarp f 'O rest plays ar 'O le in res 甘icting the distributi 'O n and p'O pulati 'O n sizes 'O f many carniv 'O re species. C 'O ntinued unc 'O nu 叫led burnin 邑and 出e spread 'O f 伽y dipteroc 創 p f 'O rest at 白e' expense 'O f mixed decidu 'O us and evergrl 田 nf 'O rest threatens the l'O ng term health health and stability 'O f 由e carniv 'O re c'O mmunity.

Current Fire Practices and Management Recommendations

S'O me 'O f the m 'O st c'O mm 'O n reas 'O ns f'O r the deliberate setting 'O f frres in s'O uth -e ast Asia Asia illclude: the c'O ntr 'O l 'O f insect pests and diseases ，clearing 'O f undergr 'O wth f'O r tra- veling veling through the f'O rest ，extracting w 'O'O d- 'O il fr 'O m Dipterocarpus species ，shifting cultiva- ti 'O n，pr 'O ducing new grass f'O r at 町acting wild ungulates ，careles.sness with cigarette butts ， matches matches and c創 npfires ， and simple mischief. At 出e present time ， the 'O fficial p 'O licy 'O f the the R 'O yal F 'O rest Department (RFD) 'O f Th ailand is t 'O prevent and st 'O p fires ，particul 訂 ly in in protected areas such as nati 'O nal parks and wildlife sanctuaries. In reality ，this policy is is '0負 .en blatantly ign 'O red and ， in areas such as Huai Kha Khaeng Wildlife Sanctuary ，日間 S are are set deliberately. Many g'O vernment 'O fficials and RFD w 'O rkers believe that seas 'O nal fires are m 'O stly beneficia l. Their experience has sh 'O wn that the f 'O rest will bec 'O me lush and green again with with the 'O nset 'O f rains ，白紙 wild ungulates will be m 'O re frequently seen ar 'O und newly burned burned areas ，and 白at the area ar 'O und their stati 'O n will be easier t 'O patr 'Ol.百1O S， every dry dry seas 'O n in Huai Kha Khaeng Sanctuary ，fires 紅 e rand 'O mly set by the sanctuary staff ， adding adding t 'O th 'O se set by the l'O cal pe 'O ple.

Research 合om this study and 'O thers suggests 出at unc 'O ntr 'O lled fi 陀 S 如 d 出e c 'O ntinued spread 'O f dry dipter 'O carp f 'O rest is n 'O t beneficial t 'O numer 'O us carniv 'O re species. Th e auth 'O r agrees with STOTT (1 988 a&b) that attempting t 'O t'Ota1 1y exclude fire 合om 血e f 'O rest is neither feasible n'O r sensible. If fires cann 'O t be c'O ntinually prevented 'O ver 血e l'O ng term ，由朗自e ground c'O ver and litter build-up resulting fr 'O m several ye 釘's 'O f 自民 exclusi 'O nc 'O uld devastate the entire f 'O rest when a fire eventually c'O mes through. An alternative policy w 'O uld be l'O w-level ，prescribed burning ，preferably early in the the dry seas 'O n when fuel accumulati 'O n is limited. H 'O wever ，alth 'O ugh this 'O pti 'O nw 'O uld prevent 曲eh 'O tter ，m 'O re extensive fires that c'O uld darnage the dry dipterocarp f'O rest ，it w 'O uld still n'O t prevent darnage t 'O the m 'O re fire-sensitive f 'O rest f'O rmati 'O ns. Such a policy does does n'O t fully address the spread 'O f dry dipterocarp f 'O rest at 由e expense 'O f the evergr 田 n and and mixed decidu 'O us f'O問 st c'O mp 'O nents 出at are S 'O essential t 'O carniv 'O res and 'O ther m 'O re C 佃'O py de 戸 ndent species such as primates. 112 112 ALAN RABINOWITZ

In In a dry tropical forest mosaic such as the study area at Huai Kha Khaeng Sanctu 釘 y ，it is felt 由at a realistic policy might be one of limited fire exclusion combined with with prescribed burning. Pr escribed burning could be done in a grid-like pattern so that most are 部釘e burned in two to three year cycles instead of annually. Are as to be managed for large ungulates could have patches burned at different times of the dry season to to allow nu 甘itious fodder to be continuously available (MOE et al. ， 1990). Fire Fire should be excluded in some of the fire-sensitive evergreen and mixed decidu- ous ous fo n:: st areas. Certain key wildlife habitats ， such as lowland riparian sites ， should be given given special protection and m 加 agement priority. The size of the area where fire exclu- sion sion is practiced must depend upon the management capability of forest officials. If possible ，fire exclusion should extend into ecotones and boundary areas so that the fire- sensitive sensitive forest formations can spread outward. Fire exclusion zones MUST be policed effectively ，otherwise the 釘 'ea to be protected is put at 加 even greater risk from an intense fire. fire. In situations where a potential fire exclusion zone c佃 not be properly patrolled 佃 d maintained ，it is better to limit fire management to prescribed burning alone. Any fire management policy must be geared to the local situation. Factors to be considered considered include: the protected status and management priority of the forest in question ， funds funds and m 佃 power available ，regional climatic patterns ，and the forest configuration. Th ecuπ'ent priority in Th ailand ， where remaining forest areas are becoming smaller and more isolated ， should be two-fold: (1) to prevent large-scale ， intense and potentiaUy devastating devastating fires in remaining forest areas; (2) to stem the encroachment of the dry diptero- carp carp forest into the more fire-sensitive yet biologically diverse evergreen forest areas that still still exis t. Some initial steps that can be immediately implemented in Th ailand include the

following: following: the establishment of fores clear 住y dep 紅 tment directives regarding fire policy ，

a coun 釘y-wide assessment of the cu 汀'ent status of fires in protected forest areas ，prohib-

iting iting the haph 位 ard setting of fires by fore 紺 Y staff ， the appointment of "日 re officers" in each each of the parks and sanctu 訂 ies ， and the education of forestry staff and local villagers about about the potential long term effects of fire to wildlife and forest communities.

SUMMARY

1. 1. Seasonal fires are mostly anthropogenic and ， in many are 儲， help to spread the fire- resilient resilient dry dipterocarp forest at the expense of evergreen and mixed deciduous fores t. 2. 2. Seasonal fires can affect carnivore mortality and stress healthy individuals whose home range contains dry dipterocarp fores t. 3. 3. Carnivore species show a disproportionately low use of dry dipteroc 訂 p forest com- pared pared to other forest types; most carnivores took refuge in wetter 釘 eas of evergreen forest during during the driest times of year. FIRE.DRYDll 司'ER OCARP FOREST. AND CARNIVORES 113

4. 4. Seasonal water sc 釘 city and a relatively low abundance of m 佃 y prey species and

企uiting 釘'ees in dry dipteroc 釘 p forest limits the distribution of many carnivore species and results in lower species densities than might exist in other forest types. 5. 5. Small carnivores 訂 'e more affected by seasonal frres and 血e resource poor dry diptero- C 鉱p fore 鉱山佃 larger wider-ranging carnivores. 6. 百le long 飽rm health and stability of the carnivore community is threatened by uncon- trolled trolled burning and by the spread of dry dipterocarp forest at 出e expense of the evergreen formations. formations. 7. 7. A policy of to 凶 exclusion of forest fires is not feasible in Th ailand. An alternative policy policy is 白紙 of prescribed burning combined with 白隠 exclusion in high priority habitats. 8. 8. Fire management policy must be geared to local situations and consider the manage- ment priority of the forest. funds and manpower available. 問 gional climatic patterns ， and 出e forest configuration. 9. 9. If fire exclusion z∞ es cannot be properly policed and maintained ，由 en it is better to limit limit fire management in 白紙 particular 釘 'ea to prescribed burning alone.

ACKNOWLEDGMENTS

1 would like to thank Warren Brockelman ， Kathy Saterson ，Sompoad Srikosamatara ，

Belinda Belinda Stewart-Cox 佃 d Susan Walker for critical review ofthe manuscript. Field assistance was provided by Susan Walker ，Saksit Simchareon. and Suwat Kaeosrisuk. Th is project was funded by Wildlife Conservation International of Th e New York Zoological Society and and was carried out in conjunction with Th e Royal Th ai Forestry Department. Additional assistance assistance and resources were provided by the National Research Council of Th ailand ， the Department Department of Biology ， Mahidol University. and Wildlife Fund Th ailand. 114 114 ALAN RABINOWI T' l

REFERENCES

BARTLE 廿， H.H. 1956. Fire ，primitive agriculture ， and grazing in the tropics. Pages 692-720 in Changing the Face of the Ea l1 h. Univ. of Chicago Press ，Chicago.

BATCHELDER ， R.B. 1967. Spatial and temporal pattems of fire in the tropical world. Proceedings of the Si. λ th Annual Ta l/ Timbers Fire Ecology CO 'l タrence ， pp. 278-297. BENDELL ，J.F. 1974. Effects of fire on birds and mammals. Pages 73-138 in T.T. Kozlowski and C.E. Ahlgren ， eds. eds. Fire and Ecosystems. Academic Pr ess ，New York. BERENSTAIN ， L. 1986. Response of long-tailed macaques to drought and fire in eastern Bomeo: a preliminary report. report. Biotropica 18 (3) : 257-262. BHUMPAKKAPUN ，N. ，KUTINTARA ， U. AND NAKSATIT ， N. 1985. The W i/ d Fauna of Huai Kha Khaeng W i/ dl 俳 Sanctuary. Sanctuary. Dep t. of Forest Biology ，Kasetsart University Report ，Th ailand. 19 pp. BUDOWSKI ， G. 1966. Fire in tropical American lowland areas. Proceedings of the Annual Ta l/ Timbers Fire Ecology Ecology Conference 5:5-22. CHRISTIAN ， D.P. 1977. Effects of 日開 on small mammal populations in a desert grassland. J. Mammal. 58(3): 423 -4 27.

GITTLEMAN ，J.L. AND P.H. HARVEY. ，1982. Camivore home-range size ，metabolic needs and 即 ology. Beha l". Ecol. Ecol. and Sociobiol. 10: 57 --6 3.

HEYER ，W. R. 1970. Field work for a study of the amphibians and 問 ptiles of the ASRCT Sakaerat Experiment Station ，Ampoe Pak Th ong Chai ， Changwat Nakhon Ratchasima. Report 27/1 ，12 ， Bangkok: ASRCT. HOWARD ，W.E. ，R. L. FENNER AND H.E. CHILDS JR. 1959. Wildlife survival in brush bums. J. Range Manage. 12: 12: 230 ー234. JOHNSINGH ，A .J .T. 1983. Large mammalian prey-predators in Bandipur. J. Bombay Nat. Hist. Soc. 80 (1): 1-4 9.

KOMAREK ， E.V. 1969. Fire and animal behavior. Proceedings of the Annual Ta l/ Timbers Fire Ecology Conference Conference 9: 160 ー207. LEEGE ， T.A. 1968. Prescribed buming for elk in northem Idaho. Proceedings of the Annual Ta l/ Timbers Fire Ecology Ecology COI ザ'e rence 8: 235-254. LEKAGUL ， B. AND J.A. McNEELY. 1977. Mammals イThailand. Assn. for Conservation of Wildlife. Saha Kam Bhaet Co. ，Bangkok ，Thailand. 758 pp. MOE ，S.R. ， PER WEGGE AND E.B. KAPELA. 1990. Th e influence of man-made fires on large wild herbivores in Lake Lake Burungi area in northem Tanzania. Afr. J. Ecol. 28: 35 -4 3. NAKHASATHIEN ， S. AND B. STEWART.COX. 1990. Nomination of the Thung-Yai Huai Kha Khaeng Wildl. グe Sanctuary Sanctuary as a UN.E.S.C.O. World Heritage Site. Royal Thai Fores 町 Department Publication. Bangkok. Bangkok. 128 pp.

RABINO W1 'π ， A. 1989. 百 e density and behavior of 1紅 ge cats 加 a 也 y tropical forest mosaic in Huai Kh a Kh aeng Wildlife Sanctuary ，百 ailand. Nat. Hist. Bull. Siam Soc. 37(2): 235 ー25 1. RABINOWITZ ， A. 1990. Notes on the behavior and movements of leopard cats ， Felis bengalensis ， in a dry

回 pical fo 問 st mosaic in 百lailand. Biotropica 22(4): 397 -4 03. RABINOWITZ ， A. 199 1. Behavior and movements of sympatic civet species in Huai Kha Kh aeng Wildlife

Sanctuary ，百 ail 飢 d. J. ofZool. ，Lo ndon 223. (11). RABINOWITZ ， A. AND S. W AL 阻 R. 199 1.百 le camivore community in a dry tropical forest mosaic in Huai Kha Kh aeng Wildlife Sanctuary ，Th ailand. J. ofTropical Ecology 7，p，制 1: 37 ー47.

ROUND ， P. 1988. Resident Forest Birds In Tha i/ and: Their Status and Conser l' ation. Int l. Council for Bird Preservation ，Monograph No. 2，Cambridge ，England. SCHALLER ， G. 1967. The Deer and the Tiger: A Study of W i/ dlife in India. Univ. of Chicago Press ，Chicago. SINGER ，F. ，W. SCHREIER ，J. OPPENHEIM AND E.O. GARTEN. 1989. Drought ，fires ， and large mammals. Bioscience 39 (1 0): 71ι722. FIRE ，DRY DIPTEROCARP FOREST ，AND CARNIVORES 115

Smπ ， P. 1984. The savanna forests of mainland southeast Asia: 釦 ecological survey. Progress in Physical

Geography 8(3): 315 ・335. STOIT ， P. 1986. Th es 戸， tial pattem of dry season fires in the savanna forests of 百 ailand. J. 01 Biogeography 13: 13: 345-358. S杭万T， P. 1 988a. Savanna forest and seasonal fire in South East Asia. Plants Today Nov /De c: 19 6- 2α1

STOIT ， P. 1988b. 百le forest as a phoenix: towards a biogeography of fire in mainland south east Asia. The Geographical Geographical Journal 154(3): 337-350. SUNQUIST ，M.E. 198 1. Th e social organization of tigers (Panthera pardus) in Royal Chitawan National Park ， Nepa l. Smilhsonian Contribulions 10 Zo ology No. 336. 98. pp. SUITHIVANITCH SUITHIVANITCH S. 1989. Ef} を'CIS 01 Fire Frequency on Vegelalivn in Dry Diplerocarp Foresl al Sakaeral ， Nakhon Ralchasima Province. M.S. Th esis ，Kasetsart University ， Bangkok. THITATHAMAKUL ， P. 1985. Vegelalion Change Along Ihe Alliludinal Gradiel in Huay Kha Khaeng Wildlife Sancluary. Sancluary. M.S. 明暗sis ，Kasetsart University ， Bangkok. 157 pp.

WALKER ， S. AND A. RABINOWITZ. In press. Th e small mammal community of a dry tropical fo 問 st mosaic in central central Th ailand. J. 01 わ'opical Ecology. WHARTON ， C.H. 1966. Man ，fire and wild cattle in North Cambodia. Proceedings ollhe Annual Tall Timbers Fire Fire Ecology Co ψ 'rence 5: 23 -6 5. WHARTON ， C.H. 1968. Man ，fire and wild cattle in Southeast Asia. Proclldings of the Annual Tall Timbers Fire Fire Ecology COI 俳'rence 8. WIL お， G .J. 1980. The birds of Salak Phra Wildlife Sanctuary ， southwestem Th ailand. Nal. Hisl. Bull. Siam Soc. Soc. 28: 101-120.

WILES ， G .J. 198 1. Abundance and habitat p陀 ferences of small mammals in southwestem Th ailand. Nal. Hisl. Bull. Bull. Siam Soc. 29: 41 -4 5.