NORBERTO SANCHEZ MEJORADA, and ABRAHAM ESCARPITA H., Director and Assistant Director, Forest Department of Celulosa Michoacán Mexico, and LOUIS HUGUET, FAO Technical Assistance Officer
It has been estimated that Mexico possesses some 10 million hectares of coniferous forests, one fifth still untouched. Despite these resources and its general industrialization efforts, Mexico still imports annually the whole of its newsprint consumption requirements and more than half of its paper pulp requirements, to a value of some $21,000,000. These quantities correspond to barely 500,000 cubic meters of local softwoods (Pinus and Abies spp) which, are perfectly suitable for the production of newsprint and paper pulp. Since one hectare of conifers in Mexico yields about 3 cubic meters per year, the systematic management of 200,000 hectares would suffice for producing the newsprint and pulp which now entail a considerable expenditure of foreign exchange to import.
Some time ago, the Mexican Government decided on expansion of the domestic pulp and paper industries, and for one project an area of approximately 600,000 hectares, half of which is stocked with conifers, was selected on which to base not only a pulp and newsprint mill but also an integrated series of forest industries. The selected area is situated around the city of Uruapan in Michoacán State and is 500 kilometers distant from Mexico City, the main consumption center for pulp and newsprint (see Figure 1).
The forests are chiefly stocked with pines, which were cut over for saw timber some 20 or 30 years ago. It would have been possible to find richer untouched forests in the more remote parts of Mexico, especially in the northwest, but preference was given to poorer forests easier to log, situated near the main lines of communication and the principal markets.
The study, implementation and administration of the Michoacán project, as it is called, have been entrusted to Nacional Financiera, S. A., which is the official bank of Mexico for promoting and financing the industrial development of the country. Since the beginning of the project, the Nacional Financiera had at its disposal the services of an FAO technical assistance officer: in 1954, this officer, together with two specially-trained Mexican forest engineers (the three co-authors of this article) set about carrying out a forest inventory of the project area. This work has now been completed, and forms the subject of this article.
The general requirements of the Michoacán project were laid down as a sulphate pulp mill with a minimum capacity of 30,000 tons per year, and a newsprint mill producing about 32,000 tons of mechanical pulp which, mixed with about 8,000 tons of semibleached chemical pulp from the other mill, could produce about 40,000 tons of newsprint annually. At the same time, it was proposed to establish a sawmill which would absorb the balance of the timber supplies left over from the requirements of the two pulp mills. Ultimately, a veneer mill was envisaged to use high-grade logs and an activated charcoal plant to use the sawmill waste and the hardwoods of the area which are generally of too poor quality for conversion to lumber. All these mills were to be fully " integrated " with one another so as to ensure the most economic use of all the wood removed from the forest.
The minimum roundwood requirements for the project were estimated to be:
|
|
cubic meters |
|
Chemical pulp 30,000 tons per year, equivalent to 5 m³ roundwood per ton |
160,000 |
|
Mechanical pulp: about 32,000 tons, equivalent to 3 m³ roundwood per ton |
100,000 |
|
Sawmill: capacity estimated at an arbitrary minimum |
100,000 |
|
TOTAL |
360,000 |
The sawmill, however, should yield about 20 percent of its intake as edgings and " waste " which could be converted into chemical or even mechanical pulp, so that the minimum volume of log supplies required, 8 centimeters in diameter and over without bark, would be only 330,000 cubic meters annually.
The forest inventory should therefore:
1. determine whether the selected forest area could actually supply on a sustained-yield basis a minimum of 330,000 cubic meters per year, and the exact allowable cut so as to determine the definitive capacity of the proposed sawmill;2. reveal the characteristics (species, average log length and diameter) of the log supplies which would be available for the various mills so as to plan their equipment accordingly, and permit estimates of the investment requirements;
3. permit the preparation of forest maps and the management plan, including prescriptions for cutting and extraction;
4. result in a clear cadastral or land ownership map and estimates of the allowable cut for each forest holding held by an individual owner, of which there were over 3,000.
It was necessary to verify rapidly that the allowable cut of the forest was at least equal to the 330,000 cubic meters required by the contemplated project. A preliminary cruise was accordingly carried out in November and December 1953.
A start was made by assembling all the available material on the forests in this area: studies, sketches, maps, triangulation points, old aerial photographs, etc. Then reconnaissance surveys were made by all available means, on foot, on horseback, by light motor lorry and by aircraft.
The whole exercise showed that there were at least 100,000 hectares of good exploitable forest, with a minimum annual yield of about 450,000 cubic meters. There was, therefore, a safety margin of 120,000 cubic meters over the required minimum annual cut, which was considered sufficient to justify the whole project as practicable, after taking into account the domestic needs of the local people.
Other results of the survey were:
1. The already existing triangulation points were marked and, taking them as the bases for further triangulation, it was possible to decide which were the other points to be determined and to plan this supplementary triangulation.2. A firm contract was drawn up with a commercial firm to carry out the aerial survey work, detailing the area to be covered, scale and type of photographs, and flying height.
3. The principal routes that could subsequently be used for carrying out the triangulation and forest sampling were ascertained and a sketch map of the forest areas was made. A rough estimate could be made of the extraction and delivery costs of the raw materials to be supplied to the selected mill site.
4. It was possible to determine fairly accurately the nature of the raw material supply in terms of species, diameter and length of logs or billets. Samples of the various species and diameter-classes were taken and tested as to which methods should be used for the production of newsprint.
5. The requirements in staff for a detailed inventory were ascertained (engineers, draftsmen, chiefs of sampling teams, workmen, drivers).
The data obtained, together with cost estimates and equipment requirements, were submitted to the National Financiera and approved at the beginning of 1954.
Ground survey
To prepare a forest map a certain number of triangulation points must first be available. Within the area surveyed there were already two such points fully determined and located by the Mexican Geographical Service; the whole triangulation of the area was based on them. Altogether 24 control points were determined, so that the average distance between two neighboring points was about 20 to 25 kilometers, which entailed working at night with a Wild T2 theodolite and luminous signals.
Despite the distance between the triangulation points, it was possible to verify that the forest map was sufficiently accurate for the use that was to be made of it by comparing the resultant map with a topographical survey made independently.
Aerial photographs
Aerial photographs were taken almost at the same time by a specialized Mexican company. The camera used was a Wild RC5a with a 21-centimeter focusing lens. The size of the photographs was 18 x 18 centimeters and the average scale 1: 15,000.
Preparation of the skeleton map: mechanical triangulation
The ground control points, that is the 24 triangulation points, were plotted at 1: 16,000 on a sheet of white drawing paper fixed on a plywood table. For preparing the map a rectangular projection system was adopted, by taking as the axis of co-ordinates the meridian 102° 00' 00" longitude and the parallel 19° 30' 00" latitude which passed almost through the middle of the forest area and, as the contact point of the plane of projection with the earth, the center of these co-ordinates. In this way, any point of the area was fully defined by its co-ordinates. For practical reasons, the area was divided from the start into rectangles of 8x16 kilometers on the terrain, which served as a basis for the map sheets, the size of which, in view of the scale, was 53.33 x 106.66 centimeters.
The next phase was to carry out the mechanical triangulation, so called because it is done automatically by means of slotted templates which can be set like a star around the center of each photograph (see Figure 2). By joining the stars formed on each photograph a semi-articulated metal network is made. These points represent the projection of the ground points located in the photographs and through which the templates pass. In each photograph, the center of the photograph and that of the adjacent photographs (called the main points) are located, as are also four points called auxiliary suitably situated. An example of the skeleton map is shown in Figure 3.
Interpretation of photographs
We will now deal with the photographs on which were marked, before the mechanical triangulation, the geodesic triangulation points, the main points and the auxiliary points just mentioned. These photographs were then studied with the aid of a mirror stereoscope (Figure 4), the boundaries between the forest areas and the nonforest areas being marked on them. Within the forest areas the boundaries between the cleared and still wooded areas were marked. Finally division into strata was carried out. The limits of each stratum were marked with a soft pencil so that the mark could be easily erased with a very soft rubber without damaging the photograph. All the strata with areas of over 2 hectares that could be recognized in the photographs were delimited. The stereogram gives an example of this classification (see Figure 5). The whole of the interpretation work was entrusted to one person, so that the bias of error should be the same. What we required of our classification was not so much to have an absolute but rather a relative value since its purpose was to divide the forest into homogeneous parts, that is within which the variations and, consequently, the intensity of sampling are slighter than in the whole of the nonstratified forest.
FIGURE 3. - Connecting the template " stare " in order to form the skeleton map
Transfer to map of details recognized in photographs
All the details indicated by interpretation of the photographs and also the usual cartographic details (rivers, villages, roads, springs, etc.) were transferred to the skeleton map. A multiscope was used to make the three points on the skeleton map coincide in the eye or mind of the operator with the same three points of the photograph to be transferred; it allows for seeing the details stereoscopically, thus automatically correcting the displacement of points brought about by differences in altitude. In this way, we obtained for the whole of the area a preliminary planimetric forest map.
Organization of sampling
On the basis of the preliminary map thus prepared and of more thorough reconnaissance surveys of the area the forests were then divided into 12 natural zones that were subsequently to become management series. This division was based on the topographical and ecological features of the zone as is usual in establishing management series: the area of each series was from 15,000 to 25,000 hectares.
The next step was to make a sampling survey of each zone. Random sampling was used with a few variants justified by practical considerations, as will be seen below. The sampling rate was 2 percent; after calculating the accuracy of the sampling for each zone, it was necessary only in the case of the two smallest and most heterogeneous zones to increase the intensity to 5 percent. The sampling units were rectangular plots measuring 20 x 60 meters.
Plot dimensions in multiples of 20 meters were chosen to permit use of a flexible steel chain 20 meters long plus a 10 meter extension, one length of the chain measuring the width of each plot and 3 lengths for its length. On rough terrain, the operator could either chain by projection or estimate the slope of the terrain, using a clinometer for checking and correcting the length of the chain by using knots fixed on the extension portion, each knot corresponding to a 5 percent step of ground slope.
The locating of the rectangular plots on the ground was carried out independently of the operator. A transparent cross-ruled paper, such as that shown in Figure 6, consisting of two groups of parallel lines was placed on the map for each series. The direction of the lines of the first group was perpendicular to the direction of the lines of the second.
The lines of the first group (horizontal in the diagram) were 1.33 millimeters equidistant on the square-ruled paper and consequently 1.36 x 15,000 = 20 meters on the terrain, while the equidistance of the lines of the second group was three times greater.
Before placing the square-ruled paper over the map, and considering the terrain to be covered, the direction to be followed by the teams charged with the carrying out of the survey was chosen. This was always either north-south or east-west (or inversely). The square ruled paper was then placed over the map but in such a way that the direction of the lines of the first group coincided with the chosen direction. The lines of both groups were then numbered and lots were drawn twice. The first drawing was made within the lines of the first group and 14 lines out of 100 (100 lines of the first group represent 2 kilometers on the terrain) were drawn, using ordinary lotto balls.
Key to Strata
|
NF. |
Forest lands. |
|
F 1. |
Lands formerly occupied by forest but now used for farming after clearings were made. |
|
F 2. |
Grazing areas. |
|
Qp-I-I. |
Mixed forest of pine and oak in which oak is the dominant species. The density of stocking of the forest is low and the average tree height between 10 and 20 meters. |
|
Qp-II-2. |
Similar to the preceding type but with medium density of stocking, the tree height ranging from 20 to 30 meters. |
|
Pq-III-2. |
Pine is the moot abundant species mixed with oak. Density is medium and the tree height ranges from 20 to 30 meters. |
|
PAQ-III-2. |
Pine, fir and oak are the species mixed in almost equal proportions. There is medium stocking and the tree height ranges from 20 to 30 meters. |
|
R |
These small-sized strata are completely covered by a young growth of pine and some fir. |
It was supposed that the lines numbered 2, 6 and 12 (marked with a thicker stroke) came out on Figure 6; then each line of the first group was considered individually and, for each line, 14 lines out of 100 (100 lines of the second group are equivalent to 60 x 100 meters = 6 kilometers on the terrain) were drawn by lot within the lines of the second group that cross it at right angles.
On Figure 6, it was assumed that, along line number 6 of the first group, lines numbered 3, 9, 13 and 14 of the second group came out. We then adopted as the southwest angles or corners of the plots situated on line number 6 of the first group, the points of intersection of that line with the lines numbered 3, 9, 13 and 14 of the second group drawn by lot. The rectangular plot was indicated on the diagram by means of hachures. This procedure was used for all the other lines of the first group which gave the square-ruled paper the appearance which it has in Figure 7, on which the forest map has also been drawn. This system of selecting plots, which is rather complicated to explain but can be carried out very rapidly, was adopted because it allows an almost fixed number of plots (7 on 3 kilometers of coverage) to be concentrated on the same line covered by the survey team, thus doing away with the need for the latter to change its direction. If the plots had been arranged without a certain amount of grouping, there would have been a likelihood of many errors in orientation and plot location.
Thus for a rectangle of 2 x 6 kilometers, that is 1,200 hectares, 14 x 14 = 196 plots of 20 x 60 meters = 0.12 hectares were delimited, the total area of which was 196 x 0.12 = 23.52 hectares, that is exactly 1.96 percent of the forest or approximately 2 percent.
Location of the samples on the map and in the field
The square-ruled paper was placed over the preliminary map and the corners of each rectangle marked on it. To find the plots on the terrain easily and quickly it was sufficient to retransfer to the photographs only an easily recognizable single point of each line obtained from the first drawing of lots. If no easily recognizable point, such as a characteristic tree, a house, the angle of a road or field, was found on this line, a nearby point was located, connected with the point itself by azimuth and distance. These were called starting points.
For practical purposes, the lines were grouped in blocks measuring 2.5 x 3 kilometers. The distance of 3 kilometers was in fact the distance that a team could cover in a single day in surveying 7 plots on a single line of the first group (first drawing of lots). The survey teams were therefore provided with photographs showing the starting points of each line of plots, as well as with the survey diagram on scale paper where the position of the survey plots was indicated in relation to the starting point. An example is given in Figure 8.
As all the distances were multiples of 20 meters and all the azimuths corresponded to one of the cardinal points, errors were reduced to a minimum (Figures 8 and 10 illustrate some aspects of this work).
A sufficient number of trees were felled in order to compile volume tables. These were double-entry tables but the heights were grouped only in three categories counted in tens of meters, in concordance with the photo-interpretation by stereoscope. This estimate was afterwards checked with the findings entered on the survey sheets (Figure 11).
Grouping samples by strata: preparing a definitive map
It has already been stated that strata had been drawn on the preliminary maps each plot could be located on the map simply by placing the transparent square-ruled paper over it, it was possible to ascertain to which stratum of the map each surveyed plot belonged. The plots were accordingly grouped by strata. After the plots had been grouped it was also possible to check the interpretation made with the aid of the stereoscope and to correct the preliminary map, which then became the definitive map.
Determination of the average hectare of each stratum
The total volume of each botanical genus, was estimated separately for each plot. By working out the arithmetical average of these findings, an initial estimate of the average volume (by botanical genus and stratum) of the plots was obtained. Then the average number of trees per plot was determined for each stratum, genus and diameter-class (in 5 centimeter categories). Thus, a second method for estimating average volume per plot was obtained.
Multiplication of the result by the factor 1/0.12 (since the area of a plot was 0.12 hectares) gave the average volume per hectare.
The same procedure was used for the average number of trees per hectare and per diameter-class.
Increment
Many samplings were made with a Pressler increment borer during the inventory of each plot. The samplings were also grouped by strata and the average number of years taken to pass from one diameter-glass to the next was determined. The increment on the average hectare was determined for each diameter-class and then for all trees on this average hectare by standard methods.
FIGURE 8. - Diagram of survey. D is the starting point while E indicates the line of the plots.
FIGURE 9. - Location of sampling unit by compass and chaining.
Projection of average values to the series as a whole
The area of each stratum was determined by the use of a planimeter and it was multiplied by the figures obtained for the average hectare, such as total volume per hectare, or per diameter-class for a single botanical genus, number of trees, increment, etc. In some cases, certain strata were not sampled on account of their small areas but, by comparison with similar strata belonging to other series of the forest zone or to comparable strata of the same series, they were assigned certain volumes and composition per hectare: in fact these presumed volumes were never more than 5 percent of the total volume of the series investigated: any error was therefore negligible.
Calculation of error
The error per stratum was computed by standard methods of statistical calculation (computing standard deviation, standard error of the average and sampling error).
Then the error in estimating the volume of the whole series was calculated. An error for the whole area of over ±5 percent was found only in the case of the two smallest and most heterogeneous zones, and it sufficed to carry out a 3 percent supplementary survey to obtain the desired accuracy. All the sampling errors computed were calculated for a probability of 0.05. Besides these sampling errors, there are the usual errors in the physical surveying: errors of measurement, errors due to the volume-tables, etc.
Cadastral map
In preparing an accurate cadastral map of the forest lands comprised in the scheme, the aerial photographs again proved very useful. A particular team leader inspected all forest holdings on the ground and marked the boundaries of each on the photographs, where they were almost always easily recognizable. He recorded the fairly numerous instances of contested ownership. The boundaries of the holdings as seen in the photographs were transferred to the map by means of the multiscope.
Finally, to estimate the growing stock on each holding, the area of each stratum contained in the particular holding was measured on the map and the areas thus obtained were multiplied by the average volume per hectare of each stratum determined as already described. The sum of the volumes of all the strata contained in each holding was taken to be the total growing stock. A similar calculation was made in regard to increment. Obviously the smaller the holding, the less accurate are the results so obtained.
Without aerial photographs, such a cadastral map of forest lands and the estimation of the growing station of each holding would have been far more costly and taken much longer to carry out.
Drafting of the inventory report
For each series, the results of the survey were written up in a report to which were annexed the forest maps and the cadastral map of forest lands. The total volume of the series per botanic genus and per stratum were given and also the area of each stratum, the composition by diameter-glasses and by stratum, etc. All the results that could be expressed graphically were given in that form in annexes. On the basis of such extremely detailed information presented in a ready reference form, the management plan for each series could subsequently be rapidly compiled.
The following is a brief summary of the inventory data:
|
|
hectares | |
|
Area covered by aerial photography: |
558,000 | |
|
Area not under forest 1: |
241,000 | |
|
Actual forest area: |
317,000 | |
|
Clearings within forest: |
45,000 | |
|
Area covered by trees: |
272,000 | |
|
of which: |
| |
|
|
a) Pure broadleaved type of little economic interest: |
22,000 |
|
|
b) Mixed conifers and broadleaved forest: |
250,000 |
|
|
Volume of standing timber on these 250,000 hectares: |
|
|
|
|
cubic meters |
|
|
a) Pine (90 percent): |
25,000,000 |
|
|
Fir (10 percent): |
25,000,000 |
|
|
b) Oak (85 percent): |
500,000 |
|
|
Alder (15 percent): |
500,000 |
|
Gross annual increment |
| |
|
Conifers |
1,000,000 | |
|
Broadleaved: |
50,000 | |
Average annual rate of increment of conifers: 4 percent.
1 Includes 227,600 hectares of farm land or pastures, 9,000 hectares under volcanic lava due to a recent volcanic eruption, 2,900 hectares occupied by villages or other habitations, and 1,600 hectares of lakes or dams.
FOREST INVENTORY OF CELULOSA MICHOACÁN ADMINISTRATIVE ORGANIZATION
Total personnel: 37 persons.
* At the end of the survey the brigade or team leaders were used In the office for the difficult work of area calculation.
The diagram shows the organizational structure at maximum staffing. Brigade and team leaders were chosen from among young men aged 18 to 25 years who had had practically no training in forestry but a secondary education and a sturdy physique. It was necessary to train them completely in the use of the compass, the chain, recognition of tree species (12 species of pine), the use of the Pressler increment borer, and so forth. After a kind of probationary period of four to five months during which almost two out of three of the trainees were eliminated, it was possible to produce an energetic and enthusiastic team of two brigade leaders, six team leaders, and assistants.
When work in the field was under way, it was possible to set the daily work norm for each type of terrain (for example, seven plots on average terrain), and to increase efficiency a bonus was awarded for each plot measured over and above the norm. To guard against the risk of quality being sacrificed to quantity, one of the forest engineers or one of the brigade leaders made periodic spot checks of the measurements recorded for completed plots.
The total cost of the inventory worked out at US$ 92,000. The cost per square kilometer was as follows:
|
|
$ |
|
Taking of photographs |
2.80 |
|
Preparation of map |
6.40 |
|
Sampling in the field |
6.40 |
|
Office work and drafting of the survey report |
4.00 |
|
Administration expenses |
2.40 |
The first two figures relate to the total photographed area of 558,000 hectares, while the last three figures relate only to the forest area of 317,000 hectares. The average cost per square kilometer under forest was therefore $29.00 and per photographed square kilometer (57 percent under forest) $16.50.
This was an intensive survey carried out not only for the purpose of acquiring a general knowledge of the resources of a region but also with a view to providing the basis for subsequent management and utilization plans. This kind of survey is always the most expensive.
If the cost of the inventory is related to an annual cut of 750,000 cubic meters of softwoods (about 75 percent of the increment) and if it is desired to pay off the costs of the survey in 10 years, the cost per cubic meter of field is 92,000/(750,000 x 10) cents, which is a low figure, even if interest is added, in comparison with the stumpage rate of timber in Mexico which is itself low, averaging approximately U.S. $2 per cubic meter (sawtimber and pulpwood).
Translated from an original French text
Courtesy: Fotoediciones " La Casa del Turista " Mérida
