F i r e M a n a g e m e n t A s s e s s m e n t o f t h e

Caribbean Pine (Pinus caribea)Forest Ecosystemso n A n dro s a n d A b a c o I s l a n d s , B a h a m a s

TH E G LOBAL FI R E I N ITIATIVE

GFI publication no. 2004-1

S e p t e m b e r 2004

C itation: Myers, R ., D. Wade, and C . Bergh, Fire Management Assessment of theC aribbean Pine (Pinus caribea) Forest Ecosystems on Andros and Abaco Islands,Bahamas. G F I publication no. 20 04-1. The N ature C onservancy, Arlington, V A .

T N C ’s G lobal Fire Initiative acknowledges the financial support of the U S D A ForestService International Programs for the development of this assessment.

For more information, contact:

Ronald MyersT N C G lobal Fire InitiativeTall Timbers Research Station13093 Henry Beadel D riveTallahassee, F L 32312850 -668-5569rmyers@tnc.orghttp://nature.org/initiatives/fire

C over Photo: C aribbean pine forest on Abaco, the Bahamas ©Ronald Myers

2

SSeeccttiioonn PPaaggee

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1O bjectives & Focus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1C aribbean Pine Forest D istribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2Pine Vegetation H istory ( Andros) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3O ther Pyric Vegetation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3

O bservations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4Fire Ecology & Fire Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4Role & Impact of Forest Exploitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5C urrent C onditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6Invasive Species and Land C onversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

Recommendations & Next Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13C urrent Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13Previous Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15Next Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17

iii

contentscontents

Objectives & FocusFrom 2-7 February 20 03, a team of fire man-agement and fire ecology experts visited theislands of Andros, Abaco, and New Providencein the Bahamas to gain insight into the fireissues facing the conservation of the C aribbeanpine forests, which cover large percentages ofeach island. The objectives for the assessmentwere to:

(1) G ather information on fire managementneeds and issues for the pine forests on thethree islands, particularly Andros and Abaco;

(2) Assess current fire regimes and discusswhether they are significantly altered fromwhat is believed to be appropriate to maintainthe integrity of the pinelands;

(3) D evelop a list of research needs and infor-mation gaps;

(4) Evaluate fire management planning andtraining needs;

(5) Identify managers to participate in trainingcourses, exchanges and mentoring programs;

(6) Identify other threats related to fire, e.g.,invasive species and climate change; and

(7)Provide Bahamian conservation managerswith recommendations and next steps.

The assessment was conducted as part of theLatin American & C aribbean FireManagement Network funded through TheN ature C onservancy by U.S. Forest ServiceInternational Programs.

The information in this report is based onobservations by, and discussions among, themembers of the Assessment Team and theirBahamian hosts during two days in the field on

1

introduction1introduction

Team Members– Chris Bergh, TNC Conservation Program Manager, Florida Keys– Frank Lowenstein, TNC Director Forest Conservation Program, NE US & Caribbean

Division– Dr. Ronald Myers, TNC Senior Fire Ecologist, Global Fire Initiative– Dr. James Snyder, Fire Ecology Researcher, U.S. Geological Survey, Big Cypress

National Preserve, Naples, Florida– Dale Wade, Fire Researcher, U.S. Forest Service (retired), Macon, Georgia

The team was accompanied by– Christopher Russell, Senior Forest Officer, Bahamas Department of Lands & Surveys– David Knowles, Assistant Agricultural Officer, Bahamas Department of Agriculture,

Abaco– Randolph Burrows, Bahamas National Trust– Ron Pagliaro, Abaco Friends of the Environment– Erin Pagliaro, Abaco Friends of the Environment

C entral and North Andros, two days on Abaco,and a brief tour of the pinelands of centralNew Providence near the airport. This reportwill focus primarily on Andros and Abaco. Theteam’s observations were limited on bothislands to broad transects along primary andsecondary roads within the pine forests. Thesetransects do not necessarily represent the con-ditions of the entire pine-dominated land-scape.

Several members of the Team (Bergh, Snyder,Myers, and Wade) are experts in the ecologyand management of the pine rocklands insouthern F lorida and the F lorida Keys (seeMyers, 20 0 0 ; Myers & Ewel, 1990 ; Snyder,1986; Snyder, 1991; Snyder et al., 1990 , Wadeet al., 1980 ). These members consider theF lorida pine rocklands to be ecologically verysimilar to the Bahamian pine forests in termsof forest structure, species composition, pineregeneration, stand dynamics, and responses to

fire. Much of this report is based on inferencesmade about the dynamics of the Bahamianpine forests based on the ecology of theF lorida pine rocklands.

This report builds on a fire managementassessment of the pinelands of G rand Bahamasand Abaco conducted in 1997 by the fire staffat Everglades N ational Park, F lorida (Segar,1997) for the Bahamas N ational Trust.

Caribbean Pine Forest DistributionWithin the Bahamian Archipelago, only four ofthe islands of the Bahamas support pinelandvegetation: Andros, Abaco, G rand Bahamas,and New Providence. Apparently, there wereonce pines on Berry Island. There are also pineforests in the Turks and C aicos, which are partof the Bahamian Archipelago, but are not partof the Bahamas nation. In the Bahamas, pineforests comprise 23 percent of the terrestrialecosystems. Andros contains 55 percent of thecountry’s pine forests. Among the four pine

2

intro

duct

ion

Figure 1. Blue hole on Central Andros surrounded by pyric Caribbean pine forest on limestone sub-strate. Sediments from blue holes provide pollen and charcoal evidence of the historical dynamics offire-dependent pine forest vegetation. (Photo: R. Myers)

islands, most of the remaining intact pinelands,approximately 350 ,0 0 0 acres, occur on govern-ment-owned “ C rown Lands” which includenational parks, national forests, freshwaterrecharge and supply areas, and unclassifiedlands.

Pine Vegetation History (Andros)Information on the long-term vegetation his-tory of Andros is available from a pollen/char-coal analysis of the sediments from C hurch’sB lue H ole, a deep, water-filled, solution hole inthe limestone bedrock ( K jellmark, 1996)( Figure 1). The analysis indicates that pyricpine forests expanded and charcoal increasedcoincidentally with the arrival of humans onAndros about 1,20 0 yrs bp. Prior to that, theclimate may have been drier, limiting humansettlement because of lack of water. The pineforests contracted around 50 0 yrs bp coinci-dentally with the introduction of Europeandiseases, and forced removal of the populationby slave traders from C uba and H ispaniola.The island remained depopulated for

20 0 years. Since then, there has been anincrease in pine concomitant with re-settle-ment and population growth. The assumptionis that there has been a dynamic relationshipbetween extent of pine forest and humanburning. The competing vegetation in thisdynamic waxing and waning of the pine forest(known locally as pineyard) is coppice—theBahamian equivalent of subtropical hammocksin southern F lorida—moist-to-dry subtropicalhardwood forest on limestone substrate.

Other Pyric VegetationC o-occurring with the pine forests are fire-maintained freshwater marshes, wet prairies,palm savannas, and salt marsh ( Figure 2).These are most abundant on the lee (west)side of the islands grading into mangrove, butare also found elsewhere. Periodic fire is need-ed to maintain the herbaceous dominance ofthese areas. Without fire, salt marshes becomemangrove, and freshwater marshes becomedominated by woody species such as coco plum( C hrysobalanus icacao) and bays.

3

introduction

Figure 2. Fire-maintained wet prairie that grades into salt marsh on North Andros. Pinelands are inthe background. (Photo: R. Myers)

Fire Ecology & Fire RegimesThe Bahamian pineyards are fire-maintained,i.e. they depend on fire for their persistenceand characteristics; a specific fire regime isrequired. A fire regime is defined as a set ofrecurring conditions of fire that characterize agiven ecosystem and include the componentsof fire frequency, fire intensity, burn severity,seasonality, and pattern of burning. The fireregime required to maintain the biodiversity ofthe pineland ecosystem can be generallydescribed as frequent (perhaps within therange of 1 to 7-10 years), low-intensity, surfacefires (i.e. fuels are on or just above the groundsurface; the fire burns through these fuels),that have little impact on the overstory or onthe soil/substrate. The fuel is probably capableof burning at any time of the year, providedthere is a day or two of drying, but a pre-dictable protracted dry season runs from aboutJanuary through May and most of the burningeach year occurs during this period. The fuel—pine needle litter, combined with grasses—mediates the nature of the fire regime. Wherepresent, scattered-to-dense shrubs and palmscontribute to fire intensity and may affect theoverstory pines.

Whether a given area burns or not depends onthe presence of fuels, whether those fuels areavailable to burn (i.e. they are dry), and anignition source. H istorically, ignitions werecaused by humans and lightning. The vegeta-tion history suggests that human ignitions mayhave been very important in creating the cur-rent extent of the pine forests. Lightning canoccur during summer-season convectionalstorms and during winter-season frontalstorms. Lightning fires probably had theirgreatest impact during the transition betweenthe drought season and the summer wet sea-

son, probably in April, May, and June. This isthe same period when human caused fireswould be most extensive. H umans have proba-bly had the effect of shifting and expanding thewindow of burning to times when lightningwas unlikely to occur and cause fires, e.g., early-to mid-dry season.

Today, the majority of fires on both islandsappear to be of human origin. Because theyoccur earlier in the dry season, they tend topre-empt most lightning season fires. Fewrecords are kept, so the role and incidence oflightning-ignited fires is not known, but light-ning-struck and killed pine trees were evident.H uman ignition sources are escaped fires fromaround homes and from agricultural burning,set fires by people hunting wild pigs, inadver-tent ignitions from people using torches tohunt for land crabs at night, other inadvertentignitions, and arson. Large portions of thepinelands are probably burned prior to theonset of the lightning season. It is likely thatlittle of the pinelands escapes burning for morethan 3 years. A huge proportion probably burnsevery 1-2 years.

The biota of the pine forest, particularly theplants, have evolved adaptations to fire. Theseadaptations include those that: (1) allowspecies to survive fire (e.g., the thick bark ofmature pines; clusters of needles that protectbuds from the heat; high, open crowns thatallow heat from a surface fire to dissipate;resprouting ability of shrubs and herbs); and(2) allow the species to respond favorably tothe fire (e.g., fire prepares the seed bed forpine regeneration by creating an open, sunny,exposed mineral substrate; it also stimulatesflowering and regeneration by seed of many

4

obse

rvat

ions 2observations

species). Animals in fire prone environmentsdevelop behavioral adaptations to fire. Forexample, the endemic Bahamas parrot onAbaco nests on ledges in solution holes in thelimestone substrate, which presumably pro-vides the nest with some protection from sur-face fires.

Under the fire regime just described and giventhe known responses of the pines, an unman-aged, unharvested pine forest would probablybe all-aged, made up of small- to mid-sizedstands or clumps of trees of the same age.These even-aged clumps or groves would rep-resent areas that at one time were open andremained fire-free long enough for regenera-tion to take place and for the trees to becomeestablished.

The effect of fire regime and changes in firefrequency on ecosystem structure are illustrat-

ed in the diagram in Figure 3. Frequent firesfavor an herbaceous ground cover. Annualfires prevent sufficient pine regeneration tomaintain the pine forest over the long term.Slightly less frequent fires stimulate pineregeneration, which maintains pine domi-nance. With an interval of several decadesbetween fires, pine regeneration is again limit-ed and favors the development of a hardwoodunderstory. Without fire, the pines will eventu-ally disappear as hardwoods take their place.

Role & Impact of Forest ExploitationThe current structure and dynamics of thepinelands on both Andros and Abaco havebeen strongly influenced by previous loggingevents. C ommercial logging began on Abaco inthe early 190 0 ’s. The last commercial harvest-ing for pulp and timber was conducted byO wen- I llinois in the late 1950 ’s through the1960 ’s. Abaco was cut before Andros. C utting

5

observations

Figure 3. Changes in vegetation structure, fire intensity, and burn severity with increasing intervalbetween fires.

ended on Abaco in 1964 when a hurricanedamaged the logging operation and the compa-ny moved on to Andros. This abrupt cessationof operations on Abaco spared a last remainingarea of old-growth pine at the north end ofLittle Abaco.

O wen- I llinois reported in 1961 that most oftheir harvesting was in second growth forestwith some even-aged stands of 50 -year-oldtrees, and other even-aged stands of 25-year-old-trees. Stand density was 196 and 180 stemsper acre respectively ( O wen- I llinois, 1961).They also reported cutting some old-growthstands where the trees ranged from 125-150 years with a density of 115 trees per acre(average diameter 8.5 inches, range 5-13 inch-es). Their report did not include stems lessthan 5 inches dbh. H arvesting left five or moreseed trees per acre. This seed source has beenlargely responsible for the regeneration of theextant pine forests. It is apparent that the for-est has been able to successfully regenerate andmaintain itself after at least two intensive har-vests over the past century.

At present, there is an active, small-scale, pineharvesting permit on Andros for furniture andother specialty forest products. O n Abaco,there was a salvage logging permit issued fol-lowing H urricane F loyd in 1998, but no extrac-tion took place.

Current ConditionsThe current condition of the pineyards strong-ly reflects the past harvesting and the frequen-cy of burning. O n both Andros and Abaco theforests are uniformly monotonous, with all ofthe trees virtually the same age. Some of theseed trees have persisted. There is little or nopine regeneration within the forest. The lack ofregeneration is due to two factors:

(1) The canopy is too dense and uniform toprovide the sunlight for successful regeneration.

(2) Where canopy gaps do occur, fires may betoo frequent in most cases to allow seedlings toreach a size to survive fires.

These beg the question: why didn’t the fre-quency of burning inhibit regeneration aftersuccessive timber harvests? There is little tosuggest that fires may have been less frequentprior to past logging events. A more likely sce-nario is that fires immediately post-loggingremoved the pine litter fuel that is the primarycarrier of fire in the pinelands. Without a pinecanopy to provide needle cast to replenish thefuel, fires were less frequent for awhile, oflower intensity, and less uniform, allowing timefor the pines to re-establish. Assuming the fireregime has not changed appreciably, this sug-gests that intensive logging every 50 years orso, with seed trees left, would continue to per-petuate the forest in a condition similar towhat it is today.

In 1998, H urricane F loyd damaged portions ofthe pine forest on Abaco. The storm damageled to abnormally high fuel loads. D uring aprotracted drought in the dry season of 1999,many of these storm-damaged areas burnedseverely, killing stands of pine of hundreds ofacres in extent ( Figure 4).

Regeneration of these severely burned areaswill likely follow the same pathway as regenera-tion following logging and fire, i.e. pine needlefuel inputs have been cut off, allowing re-establishment from mature seed trees that sur-round the burned areas. This pattern of distur-bance involving hurricanes + fire has probablybeen part of the endemic disturbance regimein the pine forests throughout their historyand may be responsible for creating relativelylarge even-aged tracts of pine.

The pine forests of both Andros and Abacohave very little regeneration. This is partly dueto the lack of light gaps. Where gaps in the

6

obse

rvat

ions

canopy occur, regeneration, ranging fromseedling to sapling to pole sized trees, is fre-quently present. Instances where this regenera-tion had recently been killed by fire are alsocommon. There is likely to be little significantregeneration and a trend toward a more naturaluneven aged forest until the forest becomesmore open. H ow detrimental the frequentfires are to this trend is unknown. It may bethat the natural development of larger canopygaps will create fuel situations similar to whatoccurred following logging, but on a smallerscale, i.e. larger gaps would lack pine needlecast and would more likely escape frequentburning until the trees re-establish. H owever,the fact that the relatively open stand of old-growth pine on Little Abaco has few youngtrees points to a regime of fire that has beenpreventing successful regeneration. If thisregime persists, there will be a gradual thin-ning of the forest, eventually creating a savan-na-like aspect.

A viable C aribbean pine forest would probablyhave trees of all ages clumped into small tomid-sized groves or stands that represent areasthat were open when regeneration occurredand escaped severe fire until the trees wherelarge enough to survive. O penings would occurwhen groves of older trees died or where hur-ricanes, drought, disease, or fire happened tokill clumps of trees. A subsequent fire wouldprepare a suitable seed bed. Abaco appears tohave more patches of regeneration than theforest on Andros. This may be due to the hur-ricane history of the island.

Anything that would limit height growth ofthe smaller pines limits successful regenera-tion. We observed that a tip moth was causingdeath of the terminal leader of many seedlingsand small trees. The longer it takes for pineseedlings to get their terminal shoots above theflame zone, the longer the young pines are sus-ceptible to fire injury and death. The incidenceof this moth is probably endemic and of little

7

observations

Figure 4. Pine stand in central Abaco killed after severe fire in 1999. This was after Hurricane Floyd-damage led to large fuel accumulations. (Photo: R. Myers)

concern; however, there may be some environ-mental factors, like excessive shade under theeven-aged stand, that create conditions favor-able for the moth.

Even though the pine stands are very uniformthroughout much of both islands, there is con-siderable variation in understory and ground-cover vegetation. These differences are proba-bly due to some combination of fire history,substrate characteristics, and depth to watertable. In places, particularly on Abaco, grassesdominate the ground cover ( Figure 5).

In other places, like the old-growth forest onAbaco, bracken fern dominates the understory.It was suggested that frequent burning mightbe responsible for the aspect dominance of theferns ( Figure 6). ( Aspect dominance meansthat a single species appears to dominate thevegetation, but no judgment can be madeabout the diversity of species of the vegeta-tion.) It has also been suggested that the fern-

dominated ground cover in the old-growthforest is species depauperate, or that speciesdiversity may be declining. This has not beendocumented. The fern dominance may be theresult of season of burn. Winter (January-March) burns in F lorida pine forests produce adominance of bracken fern. When these areasare burned later in the dry season ( April-June)grasses dominate the aspect. The shift canoccur after a single late-season burn.

Some of the differences observed in groundcover may be a function of substrate character-istics and depth to water table combined withfire history. The general consensus is thatpalms are abundant on substrates that are satu-rated with water during the wet season ( Figure7). What separates the habitat of the threespecies of palms is unknown. A lthough thesubstrate is uniformly oolitic limestone, itsgross morphology varies from sharp “pinnacle”rock to smooth pavement pitted with solutionholes. Substrates with deep fissures may allow a

8

obse

rvat

ions

Figure 5. Grass-dominated ground cover in Abaco National Park. The aspect of grasses may be dueto a regime of very frequent fire. (Photo: R. Myers)

9

observations

Figure 6. Bracken fern-dominated ground cover in Abaco’s remnant old-growth pine forest may be aresponse to frequent dormant season burns. Soil characteristics may also contribute to the domi-nance of bracken fern after a fire. Note the lack of both age diversity and regeneration in the stand.(Photo: R. Myers)

Figure 7. Pine forest on Andros with abundant thatch palm possibly indicating a high water table.(Photo: R. Myers)

greater accumulation of organic matter andclays. D eep burns may remove this organicaccumulation in some areas, affecting speciescomposition. Forests with a diverse woodyunderstory may be the result of some unknowncombination of substrate and fire history( Figure 8).

Shrub-dominated understories reflect thedynamic relationship between pine forest andcoppice vegetation. C oppice will develop any-place where pines occur that has become fire-free for a sufficient period of time. C oppicetends to occur in areas that are protected fromfire. O nce established, coppice is less likely toburn than pineland vegetation because of dif-ferences in fuel characteristics. That is, there isless fine fuel, higher dead and live fuel mois-ture, and the fuel is more compact. The mostdynamic areas of coppice are those that have anoverstory of pines with pine needle litter.These situations tend to result in more severe

fires with more damage to the pine overstory.C oppice predominates on the windward (east)side of the islands, frequently on the windwardside of rock ridges that run parallel to thewindward coast ( Figure 9). The dry seasonwinds are the consistent easterly trade winds,thus wind-driven head fires would rarely affectcoppice on the east coast. Fires backing intothe winds would tend to go out when fuelschange from pineland fuels to coppice fuels.C oppice also develops in areas pocketed bydeep solution holes, which provide moist cen-tral habitats for coppice species ( Figures 10 &11). These coppice areas can expand duringfire-free periods. C onversely, they contractduring periods of drought accompanied by fre-quent fire. There was a general observationthat there were fewer pockets of coppice onAbaco than on Andros. This would have to beverified on a vegetation map or on aerial pho-tographs.

10

obse

rvat

ions

Figure 8. Diverse shrub-dominated understory on Andros may reflect fire history and/or substrateconditions. Note young pines and dead small stems in gap. (Photo: R. Myers)

11

observations

Figure 9. Abrupt transition between pine and coppice along windward side of Abaco. Fires backinginto the wind will generally go out when they reach coppice fuels. (Photo: R.Myers)

Figure 10. Coppice-pineland transition on Andros. These zones wax and wane in response to theoccurrence of fire. (Photo: R. Myers)

Invasive Species & Land ConversionSeveral of the troublesome non-native invasivespecies in F lorida’s pine rocklands and fresh-water marshes were observed on both Androsand Abaco. They include Brazilian pepper(Schinus terebinthefolious) and several aggres-sive grasses, notably N apier grass (Pennisetumpurpureum). Burma reed ( N eyraudia reynau-diana) was not observed, but is reported asoccurring on North Andros. G enerally, theseinvasive species are only abundant in or near

rockland areas that have been converted toagriculture, and they are probably restrictedfrom the pineyards proper by the frequentburning and the lack of rock substrate distur-bance. Scattered melaleuca ( M elalueca quin-quenervia) were observed in freshwater marsh-es on North Andros. This aggressive, fire-adapted species from Australia has the poten-tial to convert the marshlands and wet prairiesto forest. It generally does not do well in pineforests on limestone substrates.

12

obse

rvat

ions

Figure 11. Pineland burn on Andros that went out when it reached coppice fuels. (Photo: R. Myers)

Current Recommendations(1) The Bahamian pineyards are the largestand most intact examples of subtropical pinerockland ecosystem in the world and they sup-port a wide array of globally imperiled and rarespecies. Several are endemic to the ecosystemand/or the Bahamas. These unique pine rock-lands are the only examples with C aribbeanpine ( F lorida’s rocklands have the southF lorida variety of slash pine [Pinus elliotii var.densa], which was once classed as C aribbeanpine). The C aribbean pine in the Bahamas isan endemic variety (Pinus caribea var.bahamensis). F lorida’s pine rocklands havelargely succumbed to urban development orhave been rock plowed for agriculture.D ifficult-to-manage remnants remain in urbansettings. Because of the small size of F lorida’spine rocklands they are susceptible to invasivenon-native plant species, prescribed burning isdifficult and limited, and in the long termmany are probably not viable. The largest andmost viable tract in F lorida is Long Pine Key inEverglades N ational Park. There are also pinerocklands on several of the F lorida Keys.Because of their uniqueness, rarity, and exten-sive intact condition, the Bahamian pineland/coppice landscape on all four islands should bea conservation priority for the Bahamian gov-ernment, Bahamian conservation organizations,and The N ature C onservancy. Any significantfragmentation of the Bahamian pinelands willlead to a degradation similar to what hasoccurred in F lorida.

(2) Except for the lack of age diversity, thepinelands on both Andros and Abaco are ingood condition. The current fire regime hasserved to keep fuel loads low and maintain anintermix of coppice. This fire regime could

probably continue for 5 or maybe 10 years ormore without any adverse effect on the ecosys-tem. In other words, due to the current struc-ture of the forest, which is the result of regen-eration after logging, there is no immediateneed to change or manage the fire regime.

(3) That said (in # 2), the fires in these forestswill eventually have to be managed to allow thedevelopment of a viable forest. More variabilityin the fire regime may also be needed to main-tain the biotic diversity of the ecosystem.

(4) In order to accomplish # 3, steps should betaken soon to develop fire management capaci-ty, as this takes a number of years to develop.

(5) C oncomitant with # 4 should be the devel-opment of a fire management plan and overallmanagement plan for the pinelands. Planningshould not precede the development of firemanagement capacity. If it does, there will be aplan but no capacity to implement it.

(6) Fire management priority should be givento the national parks on Andros and Abaco,and to the remnant old growth forest on LittleAbaco. O ther priority areas should be identi-fied, particularly the water recharge areas.

(7) Public education about fire should be a pri-ority. Most Androsians and Abaconians viewthe annual dry season fires as destructive tothe forest and dangerous to their lives andlivelihood. People need to hear about the eco-logical and economic value of their pineyards,the beneficial role of fire, and the value ofintegrated fire management. This message isparticularly important to include in school cur-ricula. C hris Russell’s “ The Bahamian Pine

13

recomm

endations & next steps

3recommendations & next steps

Forest—Anything but Barren!” poster is anexcellent start.

(8) The volunteer fire departments on bothislands need to be educated on the role of fire,and trained in using fire to reduce fuels in thepineyard-urban interface and in areas wheresmoke from wildfires is a problem. For exam-ple, the Marsh H arbor Airport on Abaco hasbeen closed at times due to smoke from fires.The incidence and quantity of smoke could bemanaged with prescribed burns.

(9) The pineyards on both islands offer a sig-nificant economic timber resource. To ensure asustainable resource and long-term ecosystemviability, an ecologically-based pine manage-ment plan needs to be developed and imple-mented in areas designated for future harvest-ing. The economic value of the forests shouldbe promoted as a strategy to limit conversionof pineland to other uses.

(10 ) Some of the pinelands on both Androsand Abaco have been rock plowed and con-verted to agriculture. O nce rock plowed, a newsoil is created that will no longer support thepineland ecosystem. The experience in F loridais that abandoned rock plowed farmland leadsto a new ecosystem dominated by non-nativespecies. Some assessment of the future threatof agricultural conversion of C rown Landsshould be made. Abandoned agricultural landsmight serve as suitable sites for pine planta-tions (of native C aribbean pine) that couldtake some of the extraction pressure off thehigher quality pinelands.

(11) Some basic information about thepinelands and the fire regime should beobtained:

From a series of aerial photographs thatprobably go back to the 1940 ’s, an assessmentshould be made of changes, if any, in the rela-

tive proportion of coppice and pineyard. H as itbeen stable, or has there been an increase inone at the expense of the other under the fireregime that has existed over that period?

Basic fire information should be obtained,e.g., number of fires, area burned, and date ordates of burn should be documented androughly mapped to get better documentationof the current fire regime and as a baseline forchanges to come under fire management.

(12) Key staff members of Bahamian govern-ment agencies, local conservation organiza-tions, and The N ature C onservancy ideallyshould have some of their time dedicated tothe conservation, monitoring, and managementof the pineyard ecosystem.

(13) Some of these people in # 12 should betrained in fire management planning and firemanagement implementation. These peoplecould be responsible for further developingcapacity within their organizations and localcommunities. In response to this recommen-dation, three individuals from Abaco repre-senting the Friends of the Environment, theMarsh H arbor Volunteer Fire D epartment,and the Bahamas D epartment of Agriculture,and one person from G rand Bahamas repre-senting the Bahamas N ational Trust, partici-pated in a week-long workshop in Belize onfire planning, fire behavior, and prescribedburning that was organized by T N C as part ofthe Latin America & C aribbean FireManagement Network. U S D A Forest ServiceInternational Programs funds covered theirparticipation. Additional training andexchanges are encouraged, including buildingpartnerships with U S agencies in F lorida thatare managing fire in these ecosystems.

(14) The occurrence of non-native invasivespecies, including melalueca and the grasses,which can strongly influence fire regimes and

14

reco

mm

enda

tions

& n

ext s

teps

thereby drastically and negatively impact biodi-versity, both directly and indirectly, should bedocumented. A plan for their eradicationand/or control should be developed and imple-mented as soon as possible to minimize costsof this effort and to maximize success proba-bility. Melaleuca has been successfully con-trolled in some natural areas in southernF lorida ( Myers et al., 20 01).

(15) Research needs include:

Fire history studies; e.g., proposal of D r.Sally H orn, University of Tennessee, to look atfire history on Abaco;

Fire regime studies, particularly season andfrequency of burn on biodiversity of groundcover and pine regeneration;

Effects of fire on Bahamas parrot habitatand population dynamics;

Relative importance of pine forest cover tofreshwater recharge and supply ( T N C ’sSustainable Waters Program); and

The potential impact of climate change onthe pineland ecosystem ( T N C ’s C limateC hange Initiative).

Previous RecommendationsO ur recommendations are consistent withthose provided by the U S N ational ParkService to the Bahamas N ational Trust in 1997.The N PS focused on fire operations needs; ourrecommendations have expanded upon thoseneeds to include the ecosystem function offire, the conservation of the pinelands values,and research needs. The N PS Action Planincluded the following:

SShhoorrtt- TT eerrmm::

Identify values at risk;

C reate defensible space around values atrisk;

Begin collection of weather data;

Begin collection of fire occurrence data;

Begin repair and/or procurement of fireequipment;

Begin consultation & coordination withcooperators (fire department, forestry,landowners);

Begin planning for response to high-riskfire conditions; and

Begin fire education and prevention efforts.

MM eeddiiuumm- TT eerrmm::

Begin prescribed burning on simple, lowrisk sites; and

Begin organization of an interagency wild-land fire management presence ( ForestWarden, wildland fire engine, volunteer fire-fighters) on each pine island.

LLoonngg- TT eerrmm::

Re-evaluate burn program and revise pre-scriptions as needed to meet objectives;

Expand prescribed burn program to includeall sites; and

Support a capable interagency wildland firemanagement crew on each pine island.

Next Steps(1) D esignate lead for T N C Bahamas to iden-tify and address fire threats and strategies/actions to mitigate those threats.

(2) Lead (identified above), working withT N C Bahamas, G lobal Fire Initiative, andBahamas partners, develops a strategy andaction plan focusing on Abaco and AndrosN ational Parks.

(3) Follow up on fire training completed in20 03 by Bahamians by identifying other needsand opportunities.

15

recomm

endations & next steps

(4) D evelop a mentoring/exchange programthat will assist in developing fire managementplans for priority sites in the Bahamas.

(5) Identify staff and partners who will focuson pineland conservation threats that are fire-related, e.g., invasives species, inappropriateforestry, and land conversion.

(6) Identify and assist individuals on Androsand Abaco to develop and implement a fireoccurrence monitoring system.

(7) Pursue opportunities to integrate trainingburns with fire regime research burns on bothAbaco and Andros.

(8) D evelop a model fire management plan forAbaco N ational Park.

(9) Transfer lessons learned in the Bahamas tohelp address fire management needs elsewherein the English-speaking C aribbean.

16

reco

mm

enda

tions

& n

ext s

teps

K jellmark, E . 1996. Late H olocene climatechange and human disturbance on AndrosIsland, Bahamas. Journal of Paleolimnology15:133-145.

Myers, R . L. 20 0 0 . Fire in tropical and sub-tropical ecosystems. Pages 161-171 In: J. K .Brown & J. K . Smith (eds.). Wildland Firein Ecosystems: Effects of Fire on F lora.U S D A Forest Service G en. Tech. Rep.R M RS- G T R-42.

Myers, R . L. & J. J. Ewel. 1990 . Ecosystems ofF lorida. University of C entral F loridaPress, O rlando, F L.

Myers. R . L., H . A . Belles, J. R . Snyder. 20 01.Prescribed fire in the management ofM elaleuca quinquenervia in subtropicalF lorida. Pages 132-140 In: K .E . M . G alley& T.P. Wilson (eds.). Tall TimbersResearch Station MiscellaneousPublication No. 11. Tallahassee, F L.

O wen- I llinois of the Bahamas, Limited. 1961.A report of the stewardship to the C rownLands O fficer of The Bahamas Islands.N assau, Bahamas.

Segar, J. 1997. G eneral Fire ManagementRecommendations for the Rand N atureC enter, Lucayan N ational Park, and AbacoN ational Park. Report on a cooperativeagreement between Everglades N ationalPark and the Bahamas N ational Trust.

Snyder, J. R . 1986. The impact of wet seasonand dry season prescribed fires on Miamirock ridge pineland, Everglades N ationalPark. SF R C -86/ 06. H omestead, F L.

Snyder, J. R . 1991. Fire regimes in subtropicalsouth F lorida. Pages: 303-321 In: H ighIntensity Fire in Wildlands: ManagementC hallenges & O ptions. Proceedings 17thTall Timbers Fire Ecology C onference,Tallahassee, F L.

Snyder, J. R ., A . Herndon, W. B. Robertson.1990 . South F lorida Rockland. Pages:230 -277 In: R . L. Myers & J. J. Ewel,Ecosystems of F lorida. University ofC entral F lorida Press. O rlando, F L.

Wade, D., J. J. Ewel, R . H ofstetter. 1980 . Firein South F lorida Ecosystems. U S D AForest Service G en. Tech. Rep. SE -17.

17

references4references