Fifteen sites were located for photo monitoring points and data collected before and after burning (Tables 1, 2, 3) using the Van Horn technique that included quantitative data and reference for each point. Each photograph is associated with quantitative data and also a location where vegetation change was expected (Bradshaw, 1993). The method provides vegetation density, vegetation height, and percent coverage on density board photos for year 1998, 1999 and 2001 (Fig 3 to 11). Photos were taken from the southern reference point facing a density board held directly over the photo point 10 meters to the north. To minimize photographic variation in successive photos over the monitoring period, a tripod was used and the height of the camera measured to 5 feet. Photos were taken with a wide angle 28 MM. lens in each compass cardinal direction.
Monitoring Transect We compared the effect of fire plus hardwood removal (transects 5, 6, 7, 8, 9, and 10), hardwood removal (transects 1, and 11) and control plots (transects 2, 3, and 4). All units were treated before the burn for exotic plant species. Eleven permanent 30 meter transects running north to south were established in the management units and data was collected before and after the prescribed burn (Tables 4, 5, and 6). Data was gathered from four 1 Mts. x 1 Mts. quadrants established every 10 meters along each transect.
Other environmental data was estimated for litter depth, litter cover, and exposed rock. The other variables were native grass, herbaceous understory, hardwood seedling, vines, dead wood, palms, hardwood rooted and exotics (Tables 4, 5, and 6). Pine Transects The location of pines along the transects were registered and their height and diameter measured during each monitoring session (Table 7).
Experimental Design Using 1998 pre-treatment data we tested restoration treatments effects with a randomized Anova analysis for selected variables such us litter, native grass, hardwoods, vines, deadwood, palms and exotics, in quadrant plots, to account for differences among procedures in plots (burn plots, hardwood reduction plus burn plots, and control plots). This analysis was done to evaluate the condition of pineland before the treatments. We tested restoration treatments effects with a randomized Anova analysis for selected variables litter, native grass, hardwoods, vines, deadwood, palms and exotics using data gathered during 1998, 1999, and 2001 combined as the covariant to compare among procedures techniques.
Prescribed Burn The units that were planned for burning were 1a, 5, 7, and 8. The morning of February 16, 1999, Florida Division of Forestry (DOF) personnel began ignition of the prescribed burn with Miami-Dade Parks Natural Areas Management staff assisting. Unit 1a was the first ignited, but it did not carry the fire well leaving the center of the unit unburned. Unit 8 carried fire throughout the unit. Early in the ignition of unit 5, a spot fire began in unit 6. DOF and Parks staff decided to burn unit 6 rather than suppress the spot fire. Units 8 and 6 burned throughout the units. Due to time constraints, unit 7 was not burned.
RESULTS AND DISCUSSION
Photo-Monitoring A series of photos collected during the monitoring sessions is provided in figures 3 to 11. Additional photos are available, if needed. Corresponding quantitative and qualitative data collected at the time each photo was taken are provided in Tables 1, 2, and 3. The dense, shrubby coverage is evident in photos taken in 1998 and supported in the corresponding data. These photos and data (Fig.12) show a significant reduction in canopy, subcanopy and shrub density after the prescribed fire. Vegetation cover showed a U shape graph. A significant reduction in vegetation covers during the first year and then slowly pattern of increase in 2001 year (Fig 13). Some authors report that fire resulted in the greatest ground cover improvement as measured by species richness and herbaceous ground cover plant density (Provencher 2001).
Non-native vegetation was not observed on any of the line transects during the monitoring sessions. We acknowledge that data collected in this fashion may not be adequate for a detailed analysis of vegetation but at the same time provide the basis for a comprehensive and reasonably thorough evaluation of the project. Monitoring Transects Data from litter depth litter cover, and exposed rock were grouped and labeled as Litter. An Anova analysis combining all variables shows that there is no significant difference in the different variables using pre-treatment data (1998) in all the designated test plots. That means that the designated plots were similar physically and biologically.
In general, comparisons of each variable showed that there is no significant difference in the amount of litter, native grass, hardwoods, vines, deadwood, palms and exotics among procedure techniques. The experimental design of comparing a set of pre-fire study to a set of post-fire within one area poses a number of difficulties for statistical analysis, particularly because lack of independence prevents the use of a temporal series of samples within one site as replicates (Hulbert 1984). Also with the adequate and appropriate sorts of replication, it should be recognized that fires are not all easy to deal with as an experimental manipulation. There are many reasons for this, first the imposition and maintenance of a fire regime for a specified block of vegetation is difficult. The weather will dictate whether to control an experimental fire at the same time. Third there is a problem of scale in relation to the area of vegetation that can be devoted to each treatment.
Comparisons of the population of a burned site with the area that escape the fire are a common approach to population studies. Also any observed differences between sites may have been there regardless of the fire. Ideally, replicate burned and control sites are needed, in which the fire treatment and the control treatment are allocated randomly to sites (Green 1979). The combined effects of all these problems suggest collecting more information of unburned areas as references sites for future studies and well replicated units with the appropriate effect of scale and balanced data.
Pine Transects The age distribution of the pine population was 18 mature, 6 saplings and 2 seedlings from which only 7 mature pines survived. (Table 7). This high mortality 74% could be explained due to the intensity of fire (peat fire) that caused mortality of adult pine, which are normally tolerant of even moderate and high ground fires.
Alexander, T. H. 1967. A tropical hammock on the Miami (Florida) limestone: a twenty-five year study. Ecology 48: 863-867. Bradshaw, A. D. 1993. Restoration ecology as a science. Restoration Ecology 1(2): 71-73. FNAI, 1998. Florida Natural Areas Inventory. Information on web site: FNAI.org. Tallahassee, Florida.