1 The Use of Fire in Silviculture Abstract: In the French mediterranean area the use of fire, practiced in the past by the farmers, is now being reactivated by the forest managers. 2 Pierre Delabraze and Jean Ch. Valette It relies on the flammabilities of the most impor- tant species and on the combustibilities of the main vegetal associations. The first prescribed burns demonstrate the sensitivity of forest trees and the role of dead material and wind. The results of these tests lead also to a better evaluation of fire risk and to organization of the clearings.

During the last century, the maintenance of - use of data gathered through these investi- chestnut plantations to make harvest easier, of gations, in order to perfect the prescribed burn- pinelands to reduce fire risk and damage, and of ing methods. pastureland to get rid of ungrazed grass, was carried out by means of small winter prescribed burns. This method was nearly completely given up FLAMMABILITY OF FOREST SPECIES as a consequence of the exodus from rural areas. In forests under the authority of the state and Principle of Measurement the departments, prescribed burning was not ac- cepted as a forestry technique. Currently the use The flammability of a vegetal sample is calcu- and even the introduction of fire in forest areas lated according to the time necessary for appear- have been banned for fear of fire outbreaks. ance of flame when the sample is subjected to a Owners are given impairments from October 15 to fierce heat radiation. April 1.

The decrease in fire danger (outbreaks and Apparatus and Method especially spread) is related to the clearing of underbrush. The four techniques: clearing by Experimental Apparatus hand or machine, use of weed-killers, use as pasturelands, and prescribed burning are scien- A radiator sends out a flux of about 7 watts tifically compared in order to present a set of per square centimeter and 3 microns wavelength. efficient and often complementary methods to the The pilot flame allows the ignition of the air-gas managers. mixture resulting from the thermic decomposition of the sample, but it does not play any part in To deal with prescribed burnings, the Mediter- this decomposition. ranean Sylviculture Center of the I.N.R.A., Forest Research Department, has developed research in order of urgency established after consultation with the forest managers. Dealing first with forecasting of fire danger, the research has led to study of the silvicultural possibilities of fire along these lines:

− flammability of the main species of forest trees, in order to determine the risks of fire outbreak and to understand their development

− combustibility of the main forest associa- tions, to define the risks of fire spread

− combustibility of the forest litter respon- sible for both the outbreak and spread of fires

1Presented at the symposium on Dynamics and Management of Mediterranean-type Ecosystems June 22-26, 1981, San Diego, California.

2I.N.R.A. Station de Sylviculture Mediter- raneenne Avenue A. Vivaldi - 84000 AVIGNON. Figure 1--Flammability measurement apparatus.

Gen. Tech. Rep. PSW-58. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 1982. 475

The Fuel (table 1) - defining the fire outbreak risks and above all their sudden aggravation and their territorial Only the flammability of the fine elements of distribution the main forest species is measured, as these elements cause the fires. - making seasonal maps of fire danger, from the forest associations maps, which should give a Experimental Checklist and Significant Data sufficiently accurate account of the flora compo- sition of the various stories. For each studied species, the monthly perio- dicity of testing is reduced to 10 days when the fire danger increases, generally in the summer Table 1--Specific flammabilities during the fire time. danger period

Picked according to very precise criteria on Studied species Flammability homogeneity, the 200 or 300 grams of green matter July August September required for a specific test are put into tight bags under partial vacuum and then placed in a Calcareous Provence cooling-box; thus the water-loss is reduced to a Tree story minimum during transport. A specific test is Pinus halepensis 4 4 4 4 4 4 4 4 4 composed of 100 basic tests made in three running Quercus ilex 4 4 4 5 4 4 4 4 4 series. Each 1-gram sample is put on the heating Quercus pubescens 5 5 5 5 5 5 5 5 5 disk. The time between the moment when the sample Shrub story is set in place to the appearance of the first Quercus coccifera flame is recorded. The test is positive if the 0 2 3 4 3 4 4 4 3 Phillyrea angustifolia 1 4 4 5 5 5 5 4 4 time is less than 60 seconds. Rosmarinus officinalis 2 3 4 5 5 5 3 3 3

Cistus albidus 2 3 3 5 4 2 2 3 2 Then for each specific test these values are Thymus vulgaris 4 3 5 5 5 4 4 4 5 calculated: Herbaceous story

Brachypodium ramosum − the percentage of positive tests - 4 5 5 5 5 5 5 5

Crystalline Provence − the mean flammation time, averaging the Tree story flammation times of the positive tests, expressed Pinus pinaster in seconds 3 3 3 1 2 2 3 3 3 Quercus suber 4 5 4 5 4 5 5 5 4 − the mean moisture content, averaging the Shrub story Erica arborea moisture contents of the four samplings, expressed 2 4 4 4 4 4 4 4 3 as a percent of the ovendry weight (24 hours at Erica scoparia 4 4 4 4 4 4 4 4 4 60ºC). Arbutus unedo 0 1 3 2 2 3 1 1 1 Cistus monspelliensis 2 3 3 3 3 2 2 1 0 The close relation between the percentage of Calluna vulgaris 3 4 3 4 4 3 3 2 3 positive tests and the mean flammation time have made a scale of flammation possible. (table 2) Table 2--Flammability marks levels as a function of the percent of positive tests and the mean Results flammation time

Species can be classified at regular intervals according to their flammability. Table 1 makes Mean flammation Percent of positive tests clear that the flammability is: (time seconds) 98 to 95 to 90 to 85 to 80 to

100 97 94 89 84 80 − grossly linked to plant physiologic activity; very low or close to null at the time of bud less than 12.5 5 4 3 2 1 1 burst, it increases with the time of new tissue from 12.5 to 17.5 4 3 2 1 1 0 lignification and reaches a maximum at the time of from 17.5 to 22.5 3 2 1 1 0 0 summer dormancy from 22.5 to 27.5 2 1 1 0 0 0 from 27.5 to 32.5 1 1 0 0 0 0 − temporarily influenced by rainfall or the air more than 32.5 1 0 0 0 0 0 moisture content; specific reactions take shape-- the pluviometry, and more precisely the amount of water stored in the soil, bring about wide varia- The forester in charge of forests threatened by tions in flammability. fires is given the decision elements to rank his intervention:

Use of the Results - on areas of species which are highly flam- mable or dangerous owing to sensitive surrounding Knowledge of the specific flammabilities leads plantations to:

476

− on bushy and low-branching plants such as Test Phases and Data Collected Pinus halepensis, with its dead lower verticils, covered with dry needles, which come nearly in The vegetation and litter are both gathered contact with the high calorific potential associa- from homogeneous and representative associations tions of Ulex parviflorus, Quercus coccifera, and on eight aligned and continuous plots, each 1 Brachypodium ramosum square meter. Before each sampling a sketch of the plant layout on the soil is drawn, along with − last but not least, when he uses prescribed a description of the phenologic stages and of burning. respective measurements and weight. The crop and all the following processes are preceded by meas- urements of the air temperature and moisture COMBUSTIBILITY OF FOREST ASSOCIATIONS content and of the wind speed and bearing. Trans- port is carried out under cover. The plants are Principle planted out on the apparatus according to the sketches. A detailed checklist states the fire First restricted to the low bush associations, ignition conditions. During the burning, the fire the study of combustibility is based on a re- rate of spread and the flame front characteristics planting of the various vegetal stories, from the are recorded. At the end of the test, the unburnt litter to the bush, on a combustibility measure- plants are measured and weighed. ment apparatus, and consists in recording the data of the combustion.

Apparatus and Method

The Combustibility Measurement Apparatus (figures 2 and 3).

The litter and collected vegetation are laid on identical surfaces on eight trucks, each 1 square meter by 25 centimeters. The moving walls act as the surrounding vegetation by protecting the fire from external agents (wind) and by reflecting the heat; a top screen simulates the tree canopy. The variable thread blade fan allows creation of wind at speeds ranging from 5 to 40 meters per second.

Fuel

The study deals with these main forest associa- tions:

Calcareous Provence Crystalline Provence Quercus coccifera Arbutus unedo Ulex parviflorus Erica raborea Rosmarinus officinalis Cistus monspelliensis Quercus ilex Calluna vulgaris

Figure 3-- Combustibility measurement apparatus (inside).

Results (table 3)

The litter plays a determining part because of the surface dead fuel. Fire has great difficulty in spreading when the litter is missing, even if the living fuel is at its lowest moisture content. Only the surface of dead fuel burns when it is bulky; when it is abundant and light it encourages the fire spread. The fire rate of spread is negatively linked with the litter moisture con- Figure 2--Combustibility measurement apparatus tent--hence the propitious consequences of a light (outside). rainfall.

477

The herbaceous story transmits the fire from Table 3--Combustibilities of the main forest the litter to the shrubby story. Its phenologic associations stage and its abundance affect fire spread more by its high flammability than by the released energy. The graminaceae that dry out in summer are the Forest Associations Combustibility more dangerous; some of these turn green again Calcareous Provence after the first storms of late summer. Quercus coccifera Good in summer, poor in autumn The rain brings about a quick change in com- Ulex parviflorus Excellent in summer, bustibility. It increases the moisture content of existent in winter the litter and leads to a new start of the vegeta- Rosmarinus officinalis Poor tive activity. In summer, 50 percent of the tests Quercus ilex Similar to Q. coccifera in which fire covers the 8 meters of the apparatus Crystalline Provence are in samples taken more than 10 days after the Arbutus unedo Poor last rain; on the other hand, 85 percent of the Erica arbora Excellent in summer, tests whose combustion stops within the first Cistus monspelliensis Excellent in case of meters of the apparatus have been on samples taken summer drought less than 10 days after a rain. Calluna vulgaris Good in summer

Wind speed and rate of fire spread are posi- tively linked to one another. The increase of the flame length and the width of the flaming area the flaming front incline on the combustible explain it. A fuel bed (86 x 58 x 6 centimeters) rests on Fire spread is linked with the plant structure: scales (15.00 + 1.5 grams). Eight thermocouples for the same biomass, an effective low bulk (Nc-Na), one above the other, and a fluxmeter are density increases combustibility owing to the connected to recorders. A photographic apparatus large heat-absorbing surface and to the possible records the combustion. ventilation inside the fuel. Studied Species

Use of the Results Material is harvested in autumn from branches of Pinus nigra Arn. ssp laricio, Pinus pinaster The combustion of the low story vegetation is Soland, Pinus halepensis Mill, Cedrus atlantica correctly modelled on the apparatus: Manetti.

− the flame front specific energies are calcu- Fuel lated according to Byram's formula from the van- ished combustible mass, its moisture content, fire It has these characteristics: rate of spread, and knowledge of the heat release − surface/volume ratio of the leaf or the − the energy radiated by the flame front is needle, expressed as inferred from study of its development and of the measured temperatures (adaptation of the Stephan- 2 2 (+2) =or = Blotzmann formula) e e

− also, the biomasses and the laws of their where e = thickness of leaf or needle development are drawn from these tests. − initial mass

DEAD MATERIAL COMBUSTIBILITY − moisture content, expressed as a percent of ovendry weight Principle of Investigation − dry material initial mass Currently, dead fuel combustibility is only studied on leaves or needles picked on trees, − bulk density, , in grams per centimeter desiccated, and then laid out by hand in regular layers. Results

Apparatus and Method Dried needles of Pinus nigra laricio

Experimental Apparatus (figure 4) = 4.7 mm-1 = 1.8 10-2 g.cm-1

478

All the combustion data of the tests made in Dried needles of Pinus halepensis horizontal position are positively related to the initial ovendry mass from a threshold of steady = 6 mm -1 = 3.3 .10-2 g.cm-3 rate of spread on 3.6 t/ha. The most accurate fittings are linear and the correlation coeffi- The fire rate of spread threshold is close to cients are superior to 0.9. 1.8 t/ha. All the data are related to the initial ovendry mass, yet the slopes of the regression When the fire moves upwards, from a 20º slope, lines are significantly lower than those of the the fire spread laws are totally altered. The two aforementioned pine trees. fire spreads up the slope very fast, burning only the upper layer of the dead fuel. Dried needles of Cedrus atlantica

The tests made with fires moving downwards give = 33.10-2 g.cm-3 results which can be compared with those made on a flat position, yet the slopes of the regression Achieving a continuous carpet requires the lines are lower. equivalent of 5 t/ha. Even for the highest values the fire does not spread.

Dried needles of Pinus pinaster HIGH CALORIFIC VALUES, ASH AND MINERAL CONTENTS = 2.7 mm -1 = 2.10-2 g.cm-3 The data allow an accurate estimation of the The characteristics of the combustion of this energy released during the combustion. dead fuel are akin to those obtained with needles of Pinus nigra laricio. The slopes of the regres- sion lines are not significantly different whereas The High Calorific Value the ordinates at the origin are slightly superior. Determining this allows us to calculate the potential energy of forest combustibles and the maximum heat release. Measured every month ac- cording to the classical calorimetry method, the H.C.V. are comparatively lower during growing time than after lignifying time. The H.C.V. of Erica arborea is higher than the values given for most of the tree species (table 4). The H.C.V. of mediterranean species are generally superior to those of more northern species. As for the building material, no link between the H.C.V. and the specific flammabilities has been clearly shown.

Ash and Mineral Contents

Determined every month by decomposition at 450°C in a muffle furnace, the ash contents are low during growing time and higher after ligni- fying time. The ash content of the Arbutus unedo leaves is double that of the ash content of the Erica arborea leaves, whose Ca content is very similar to the Ca content of the Ulex parviflorus shoots (table 4).

The study did not show any clear link between these contents and the characteristics of combustion.

THE PRESCRIBED BURNINGS IN FRENCH MEDITERRANEAN AREA

Widely used in the past, the prescribed burning method is nowadays only used by a few shepherds in Corsica and some peasants in the Maures and Cevennes Mountains. Figure 4--Diagram of the apparatus for measuring the dead material combustibility.

479

Table 4--Specific high calorific values, ash and mineral contents

Fuel Contents H.C.V. Ash P K Mg Ca (kJ.g-1) (percent ovendry weight)

Leaves of Erica arborea 24.0 2.42 0.07 0.49 0.23 0.30 Needles of Pinus halepensis 22.2 3.20 0.11 0.50 0.15 0.63 Shoots of Ulex parviflorus 20.9 2.34 0.08 0.62 0.13 0.37 Leaves of Arbutus unedo 21.0 4.15 0.10 0.61 0.27 0.99 Leaves of Quercus ilex 20.3 4.05 0.09 0.50 0.15 0.96 Leaves of Quercus coccifera 20.0 4.20 0.07 0.53 0.14 1.22

Aims Results (table 6)

The tests on prescribed burnings are intended These tests being on the whole quite recent, to define the rules of use according to the cli- the results are only provisory. matic, relief, and soil conditions and to improve knowledge of the reaction of forest tree species From the first investigations, the following to the various types of prescribed burning. points can be noted:

− large damages to the tree story in tests 3, 8 Plotting Out and Method and 9, which were made at the bottom of the slope on account of an adverse wind Each test area is delimited by a surrounding cleaned area, gridded with posts to make marking − satisfactory aspect of tests 4 and 7, ignited and measurement easier to determine. at counterslope, and of tests 1, 2 and 6, which were made against the wind: the low stories are − importance, nature, and qualities of the dead well burnt back and trees are spared fuel and of the various vegetation stories − the high combustibility of Quercus ilex and − measurements of the trees where thermosensi- of Pinus halepensis foliage and the relative tive plates are set endurance of needles of Pinus pinaster.

− microclimatic conditions Prescribed burning can only be correctly carried out − fire ignition conditions, flame front spread, development of flames on trees, and reactions of − if the wind direction is well established and the various species if its speed is low and steady

− aspect of the ground and conditions pre- − if the dead fuel or herbaceous layer is vailing there after the test. continuous enough to guarantee a steady rate of fire spread Later, the restoration and new setting up of vegetation will be observed with interest − if the fuel on the ground is dry enough (table 5). − if the low verticils of trees are more than 2 The prescribed burns are generally made against to 4 meters above the low story. the wind or at counterslope.

Table 5--Scale for obvious fire damages Conclusions

Obvious damages Scale These first tests show that planted firebreaks No damage 0 or trimmed bushes of Quercus pubescens covering a graminaceous grassland can be maintained, that Over 2/3 of the initial leaf mass are green 1 there are maintenance problems in planted fire- breaks where there are no resinaceous species, From 1/3 to 2/3 of the initial leaf mass are green 2 owing to a scarce litter, and that there are limits to the use of prescribed burning in Less than 1/3 of the initial leaf mass is green 3 forested areas as soon as the mattoral (maqui and Some green leaves 4 garrigue) are fully grown. No green leaves 5

480

PUBLICATIONS OF THE CENTER ON THESES SUBJECTS Doat, J.; Valette, J. Ch., 1981 - Le pouvoir calorifique superieur d'especes forestieres Caramelle Ph., Clement A., 1978 - Inflammabilite mediterraneennes. High calorific value of et combustibilite de la vegetation mediter- mediterranean forest species. Ann. Sci. For., raneenne. Flammability and combustibility of to be published. the mediterranean vegetation. (3rd year thesis Valette, J. Ch.; Clement, A.; Delabraze, P., E.N.I.T.E.F., July 1978, 158 p.) 1979 - Inflammabilite d'especes mediter- Delaveaud, P., 1981 - Le feu, outil sylvicole. raneennes. Flammability of some mediterranean Utilisation pratique des donnees de combusti- species. (Research note 79-3, Mediterranean bilite. The fire as a sylvicultural technique; Sylviculture Center, 39 p.) practical use of combustibility data. (3rd Valette, J. Ch,; Vannier, G., 1981 - Inflamma- year thesis E.N.I.T.E.F., July 1981, to be bilite estivale des principales especes published.) forestieres de Provence calcaire. Summer flammability of the main forest species of calcareous Provence. To be published.

481

Table 6 --First tests of prescribed burnings

482