Approach and methodology

Survey team

Team Leaders: William Milliken (EDC, RBGE) and Bruce Albert (CCPY, IRD)

 

·         Dr Bruce Albert (anthropologist, Institut de Recherche pour le Développement) 

·         Rogerio Duarte (Masters student in anthropology, Universidade de São Paulo)

·         Dr Francois-Michel Le Tourneau (geographer and remote sensing specialist, CNRS-CREDAL)

·         Dr Maria de Lourdes Ruivo (soil scientist, Museu Paraense Emílio Goeldi)

·         Dr William Milliken (ethnobiologist, Edinburgh Development Consultants and Royal Botanic Garden Edinburgh)  

·         Edinelson Pereira (indigenist, responsible for Comissão Pró-Yanomami’s environmental programme)

·         Manoel da Vera Cruz Leal Bezerra (indigenist, URIHI health programme)         

·         Raimundo Yanomami (Watoriki theri)                                                 

·         Henrique Yanomami (Watoriki theri)

Background research and preparation

Background research was focused on gathering existing data on the Homoxi region including topographic maps, remote-sensing images and survey reports (botany, zoology, hydrology, anthropology etc.).  Standard plasticised base maps (based on satellite images) were prepared for use in the field.  Representatives of the Yanomami communities in Homoxi were consulted by CCPY staff in Boa Vista, and arrangements were made to bring two Yanomami men from the lowland village of Watoriki along with the team.[1]

Fieldwork and methods

Fieldwork was undertaken between 05 and 24 April 2002, which was shortly after the beginning of the rainy season. The team was based at the URIHI health post at Homoxi.

Botany

Qualitative vegetation descriptions were made in both disturbed and undisturbed environments, supported where appropriate by specimens and by photographs.  As broad a range of habitats as possible was visited, following identification of key sites during initial reconnaissance of the area and analysis of satellite images.  In addition to point sampling of vegetation and general observations,  detailed descriptions were made of the vegetation along two line transects across the areas most affected by gold mining, and photographs were taken of the vegetation from the air.[2]  Where possible the locations of all vegetation sampling points were recorded using a handheld GPS.  Two hundred and sixty pecimens (LH series) were collected as unicates (i.e. one of each) and preserved in 70% ethanol (Schweinfurth technique).  These were subsequently dried at the Museu Integrado de Roraima (Boa Vista) and at INPA in Manaus, and were identified as accurately as possible in the INPA herbarium.[3]  This material remains in the INPA collection.  A complete list of the plants recorded in the region is provided in Appendix 2, together with data on their habitat distribution and collection status.

Digital photographs of the vegetation were compiled into interactive ‘guides’ using html files.  These can be viewed on the CD-Rom accompanying this report.

Soils and hydrology

Soils were sampled over a wide area and range of habitats, the sites initially having been selected on a physiographic basis.  Observations and soil collection were undertaken in small trenches 50cm deep, with additional sampling to a depth of 90cm using an auger.  Morphological characteristics such as texture, structure, porosity and transition (described and classified following Lemos & Santos, 1984) and colour (compared with Munsell chart, 1975).  The gravel substrates were sampled at depths of 0-20 and 20-40cm, with the exception of a few samples that included 40-60cm samples.

All samples were air dried, broken up manually with a wooden roller and divided into four sub-samples.  Approximately ¼ of the sample from each horizon was sieved in a sieve with diameter <0.002 to 2mm, in order to assess granulometric composition (Embrapa, 1997).  Standard chemical analyses of soils <2mm were undertaken in the laboratory at the Museu Goeldi (Belém), following the methodolody proposed in the Manual de Métodos de análise de solos (Embrapa, 1997).

Table 1: Textural and chemical procedures used for soil samples from Homoxi

Analysis

Method

Texture (sand, silt and clay)

Gravimetry

PH in water

Potentiometry

Organic Carbon (C)

Walkley-Black

Total Nitrogen (N)

Kjëldahl

Exchangeable Phosphorus

Colorimetry

Exchangeable Calcium (Ca) and Magnesium (Mg

Atomic Absorption Spectrometry

Exchangeable Sodium (Na) and Potassium (K)

Flame Spectophotometry

Exchangeable Aluminium (Al)

Atomic Absorption Spectrometry

Exchangeable Hydrogen (H)

Atomic Absorption Spectrometry

Exchangeable Cations Sum (S)

Reckoning (Ca+Mg+K+Na)

Capacity Exchangeable Cations(CEC)

Reckoning (S+Al)

Saturation Exchangeable Cations (V)

Reckoning (CEC–S/Al)

Saturation Exchangeable Aluminium (m)

Reckoning

Hydrological observations were limited to mapping of water bodies and water courses from satellite imagery, direct observation of the their nature on the ground and sampling of their water (man-made lagoons and Rio Mucajaí).  The water samples were analysed at the Museu Paraense Emílio Goeldi. 

Remote sensing

On account of difficulty of access and almost permanent cloud cover, the Serra Parima region is very poorly provided for in terms of cartographic resources, particularly large scale ones.  PROJETO RADAM provided the only exhaustive cartographic cover[4] in the 1970s using sideways-looking airborne radar, providing the following types of products:

·        Thematic maps at 1:1,000,000 scale (vegetation, geology, soils etc.).  This scale is too small for localised studies but provides useful regional-scale reference material.

·        Topographic maps at 1:250,000 scale prepared from the radar images enriched with supplements corresponding to roads or primary agglomerations.  In the case of the Serra Parima these supplements are almost non-existent, as well as dated (the maps were produced in 1980 and have not been revised).  In addition, the use of radar images for defining topography in areas of high relief such as the Serra Parima tends to result in certain geometric inaccuracies.

To these resources one can add the 1:250,000 ‘planimetric maps’ which were produced in 1980, showing the principal water courses and relief uniits (without guarantee of precision), and a topographic map of Roraima, edited in 1995, at the scale of 1:1,000,00.  This last map, probably based on the planimetric map (among others), is regarded as one of the most reliable sources of information on the region - in spite of its small scale.

All these documents were scanned (for the Homoxi region) and inserted in a GIS system.  The main cartographic elements were vectorised, providing one of the sources of information used for the research.  It should be noted that it was necessary to adapt these maps in order to make them compatible with the finer-scale satellite images of the region.  In all cases where the satellite image data were interpretable, these were treated preferentially.  In areas of cloud, however, where satellite data could not be interpreted, the cartographic data were used.

For larger-scale maps (for general use and localisation), numbered small-scale geographic views were used, such as those produced by ANEEL (Agence Nacionale de l’Ectricité) at 1:3,000,000 or those of the Diagnostico Socioambiental da Amazônia prepared by IBGE in 1996.  In these two cases the generalised large-scale views are not appropriate for local-scale cartography.

Given the limitations of cartographic resources and the overall aims of the survey (mapping the areas affected by gold mining), it was decided to use remote sensing images as the base maps for the fieldwork.  However, there were numerous obstacles to the acquisition of these images.

Use of radar images – On account of the prevalence of cloud cover in the Serra Parima region, particularly around 10.00 when most of the heliosynchronous satellites pass over the region, the use of radar images was proposed on account of their ability to ‘see through clouds’.  Two SAR images from the European satellites ERS-1 and ERS-2 were acquired from the European Space Agency (ESA) thanks to the Laboratoire Géomatériaux of the Université de Marne la Valée.  With these two views on different dates we hoped to be able to evaluate the changes in ground cover since the abandonment of mining activities in 1990.  Unfortunately, however, although the images were of a high technical quality, it was impossible to use them for this purpose.  Radar tend to exaggerate considerably the effects of relief, and slopes orientated towards the camera are over-exposed to the extent that no further information can be taken from them.  Conversely, the rest of the terrain (facing away from the camera) is in shadow and likewise yields no additional data. 

Figure 1: Radar image showing the Homoxi region (in box)

Thus the radar images were not used to the extent that we had hoped.  However, avenues of research do exist that may permit their use, either by the use of software to reduce the effects of relief or with the acquisition of new images from other sources with more appropriate wavelengths and angles of incidence, or even with more than one polarisation.   It should be noted that the frequency of image acquisition by ERS satellites in the region is important, and could provide a temporal sequence with an annual image between 1992 and 2002, which is far from the case with optical images.  Thus, with the use of appropriate interpretation methodologies, a precise chronological study of the mining sites (placers) could be carried out between these dates.[5]

Images acquired:

ERS1 SAR orbit 6983 frame 3555 track 325 from 15/11/1992

ERS2 SAR orbit 24728 frame 3555 track 325 from 12/01/2000

ERS2 SAR orbit 25000 frame 3555 track 96 from 31/01/2000

Use of optical images – Due to the difficulty of using radar images and the lack of reliable cartographic data, optical views constituted the main source of images for our GIS.  However, as has been mentioned, the cloud cover substantially reduces the choice of usable images for this analysis.  No suitable images were identified in the SPOT catalogue (these have the best resolution for the type of work envisaged in the study area).  Two Landsat images (one TM from 1989 and one ETM+ from 1999) were identified which were usable for the Homoxi region (it appears that no other ones exist for the 1986-2002 period!), the first from a catalogue of images acquired by the Cuiaba station and the other from the Landsat 7 catalogue.[6]  The limitations of cloud cover would also have to be borne in mind if considering the possibility of refining the mapping with the use of very high-resolution images such as Ikonos.

The images were geo-referenced using the coordinates provided with the ETM+ image, whose remarkable accuracy was confirmed on the ground by the acquisition of GPS points at easily recognisable sites such as the Homoxi airstrip etc.  The 1989 image was overlain on the 1999 image and cartographic data were corrected, as far as necessary, on the basis of this image (which constituted our reference).

Images acquired:

Landsat 5 TM 001/058 du 28/02/1989

Landsat 7 ETM+ 001/058 du 15/11/1999


Figure 2: Landsat image of the Homoxi region, 1989


Zoology

The depth of the zoological component of the survey was limited by the relatively short period in the field, which was insufficient to make significant direct observations of the fauna.  Lists of species (focusing on those of economic importance) were made during discussions with Yanomami hunters, the names initially recorded in the Yanomami language.  These were then identified as far as possible with the help of the following illustrated field guides.

Birds                                      Schauensee & Phelps (1978)

Mammals                               Emmons (1990)

Fishes                                   Santos et al. (1984)

In addition, a few specimens of small animals (e.g. prawns, fishes, tadpoles etc.) which were brought in for food by Yanomami women were preserved in 70% alcohol and taken to INPA (Manaus) for identification.  These have been incorporated in INPA’s zoological collections. 

Socio-economic survey and ethnobiogy

Data were collected from both male and female Yanomami informants using semi-structured interviews conducted in the Yanomami language.[7]  These included information on the following subjects:

·         Effects of gold mining on the livelihoods of the Yanomami communities

·         Effects of gold mining on the environment, and on species of significance

·         Recuperation and other changes since the termination of mining activities

·         Environmental and economic priorities of the Yanomami communities

·         Options for implementation of a recuperation programme

Additional data were collected with the help of mapping activities, conducted in groups using large sheets of paper and coloured pens.  These included information on place names, past and present land use and occupation and the distribution of key resources, habitats and species.

Initial ethnobotanical data were collected by free listing of species within specific use categories (e.g. food plants, medicinal plants, plants for construction, poisons, constsituents of hallucinogenic snuff, environmental indicators etc.).  Plants were listed by their Yanomami names, and efforts were then made to collect samples of these species with the help of knowledgeable individuals.  In the course of collecting these species additional data and samples were taken as useful species were encountered by chance.  Upon return to the village, data were cross-checked with other knowledgeable individuals using the plant material as a reference point.

Entomology

Bees were collected primarily from the open areas (gravels, airstrips, river bank and vicinity of the Post) using a hand net.  Where possible collection was undertaken jointly with Yanomami men, and notes were made of their Yanomami names and other characteristics.  Specimens were pinned out and dried, and taken to INPA for identification.[8]

Tree planting

If an environmental/ economic recuperation project in the Homoxi region is to be successful, the active participation of the Yanomami will be crucial at all stages.  In addition to involving the communities in the survey and the initial decision-making processes, it was therefore deemed appropriate to begin practical activities that would initiate the process of environmental recovery and economic benefit.[9]  Experimental fruit tree planting was therefore carried out by Edinelson Macuxi in collaboration with several members of the Tirei community (around 16 individuals, of both sexes) whilst the rest of the team was carrying out other fieldwork.

Four sites were chosen for planting during discussions with the Tirei theripë.  In the first area, a patch of gravel located between two lagoons close to the Post and the maloca, 341 cashew seeds of various varieties (obtained in Boa Vista) were planted.[10]  The men dug concentric circles of holes, starting on the perimeter of the gravel area and finishing in the centre.  Organic waste from the maloca was put into each hole to assist germination, and the women then planted the seeds and filled the holes.

The second chosen area was also very close to the Post, on the left bank of the Mucajaí.   Seeds were planted using the same techniques, but instead of an arrangement of concentric circles they were planted randomly by the Yanomami.  Seventy seeds were planted at this site, and a further 322 (also randomly) in a gravel area on the other side of the river.

The last site was situated at the end of the Homoxi airstrip, close to a large lagoon.  300 seeds were planted in random fashion, without the addition of organic matter to the holes.  These sites will be monitored over the following year by the Yanomami and by CCPY staff.

In addition to the cashew seeds, smaller-scale planting was carried out in a number of locations.  Hundreds of water-melon seeds, for example, were planted in the main Yanomami garden, and several hundred bacaba (Oenocarpus bacaba) seeds were scattered in the secondary vegetation at the base of the bank below the Post and the Tirei maloca.  These seeds had been brought by the Yanomami from Wiramapiu and were in the process of germinating.[11]

 

Figure 3: Planting cashew seeds

Figure 4: Inspecting germinating seeds

Figure 5: Mapping resource use

Figure 6: Cataloguing hair samples

Figure 7: Mapping resource use

Figure 8: Field observations

Figure 9: Collecting data with the Yanomami

Figure 10: Ground truthing satellite image

Figure 11: Robi collecting botanical specimens

Figure 12: Antonio from Yaritha distributing exchange goods at the Post

Figure 13: Marking satellite points

Figure 14: Collecting soil samples

Figure 15: Collecting bees

Figure 16: Robi, who provided much of the botanical information for the survey

Figure 17: Crossing the swollen Mucajaí after rains

Figure 18: Francisco accompanying fieldwork in the forest

 

 



[1] These two men, Raimundo and Isligue, were already well known to several of the team members.  The purpose of their inclusion was primarily to facilitate discussions with the people of the Homoxi communities, and to build confidence rapidly.  This proved to be very important, particularly in the context of allaying the communities’ concerns over soil sampling (which can apear very similar to gold prospecting).

[2]  In addition to photograhs taken on the flights into and out of the Homoxi airstrip, images were taken from a specially commissioned overflight of the area following a pre-ordained course.

[3] Given the short time available it was necessary to identify specimens by direct comparison with named material rather than with taxonomic keys and literature.  As a consequence, particularly in sections of the herbarium where specimens have not recently been curated by a specialist, there may be minor inaccuracies in the names.  However, these are unlikely to be significant.

[4] Mapping operations were undertaken as part of the survey of mineral resources carried out by the military in the 1980s, but the results are not accessible to the public nor to research organisations.

[5] The chronology of the placers is much more complex than a general decline in 1990.  In spite of the operations conducted by the Federal police at this time, the last expulsions of garimpeiros from the Homoxi region did not take place until 1998 (and in some parts of the Yanomami area gold mining is still going on).  Furthermore the continued activity of garimpeiros after 1990 was not only focused on pre-existing sites, but also involved the opening of new clearings.  A precise chronological study, such as has been conducted on the Serra Tepequém plateau, would be of considerable interest.

[6] Purchased by CCPY.

[7] The orthography used to transcribe the Yanomami words used in this article (in italics) generally follow the standard IPA usage. The Yanomae dialect spoken at Homoxi has seven vowels (i, e, i, ë, a, u, o) and thirteen consonants (p, t, k, th, hw, s, x, h, r, m, n, w, y). Of the vowels, only two are represented by non-IPA standard symbols: i refers to a close/tense, high central unrounded vowel (which has no corresponding sound in English) and ë which refers to a mid central vowel (similar to schwa /q/ vowel in the English word cut). Of the consonants, th refers to an aspirated, valveolar stop (similar to the initial consonant in the English word tip), hw refers to a labialized, voiceless glottal fricative (similar to the initial sound in the word where for some dialects of English), and x, following Portuguese orthography, refers to a voiceless palatal fricative /ò/ (similar to the initial sound in the English word ship). The palatal semiconsonant/semivowel /j/ is represented by y, according to common American phonetic usage.  See Appendix 1 for further details.

 

[8] The specimens were given to Alexandre Coletto da Silva.

[9] In projects such as this the inevitable delays between initial reconnaissance/ discussions and initiation of the project can lead to disillusionment and even mistrust in the communities.  Early initiation of intervention activities can help to prevent this.

[10] The choice of seeds was partly influenced by seasonal availability.  Cashew grows well in the drier savanna (lavrado) areas in the east of Roraima and may survive the relatively harsh environment on the gravels.  Its fruits are appreciated by the Yanomami. 

[11] Many of these were rapidly destroyed by rodents.  This emphasises the need for construction of a seed nursery for the establishment of saplings..