International Journal of Geobotanical Research, Vol. nº 6. 2016. pp. 27-36

The Place of Geobotany in Geology

(1) Beneah D. ODHIAMBO

(1) University of Venda, Private Bag X5050, Thohoyandou, 0950; Limpopo Province, South Africa. Tel: +27159628587, Mobile: +27607786672,E-Mail: [email protected]; [email protected]

Abstract: Geobotany is the study of as related to the geological substrate. In this article, the term geobotany is defined as it is understood and used among the geological fraternity. It is the visual survey and analysis of vegetation in order to discriminate geological differences in the subsurface. This chapter demonstrates the place of geobotany in geology by giving a reflective historical perspective on the use of plants in geology. Geobotany is significant in exploration of mineral resources and is of strategic importance in geologic mapping in vegetated terrain. For this purpose, the uses of plants in mineral exploration, and the entry levels at which plants are used to discriminate geological differences in the landscape are presented. These entry levels include the use of plants for general geological mapping (regional geobotany) purposes, and for specific identification of mineral deposits (indicator or target geobotany). How mineral uptake influences the spectral reflectance characteristics of plants (spectral geobotany) is presented in the last section of the chapter as a geobotanical technique that uses remote sensing technology. This historical review of the place of geobotany in geology, underscores the genesis and paradigmatic shift in research and development through which the discipline has evolved over time.

Keywords: Geobotany, element uptake, spectral reflectance, bioremediation, Kenya.

Introduction research and development through which the discipline has evolved over time. ‘Rocks produce soils in which plants grow’. Geobotany is the study of plants as related to the Definition of Geobotany geological substrate. In this article, the term geobotany is Geobotany encompasses the linkage between geology defined as it is understood and used among the geo- and plants; and specifically how plants as an interphase logical fraternity. It is the visual survey and analysis of are or can be used in geological studies, for example in vegetation in order to discriminate geological differences mineral exploration, in biogeochemical studies, in me- in the subsurface. This chapter demonstrates the place of dical geology and in remote sensing studies, to mention a geobotany in geology by giving a reflective historical few application areas. perspective on the use of plants in geology. Geobotany is In this chapter, geobotany is defined and used from a significant in exploration of mineral resources and is of strategic importance in geologic mapping in vegetated geological point of view; it is the science that deals with terrain. For this purpose, the uses of plants in mineral the visual survey of vegetation in order to identify geolo- exploration, and the entry levels at which plants are used gical differences in the landscape (Malyuga, 1964; Rai- to discriminate geological differences in the landscape nes and Canney, 1980). This visual survey often includes are presented. These entry levels include the use of observations of physiological modifications in plants, plants for general geological mapping (regional such as dwarfism, morphological deformities, and chlo- geobotany) purposes, and for specific identification of rosis (Kovalevsky, 1987). These teratological (Raines mineral deposits (indicator or target geobotany). How and Canney, 1980) changes can occur in individual mineral uptake influences the spectral reflectance plants or whole communities. In this regard, geobo- characteristics of plants (spectral geobotany) is presented tany is also concerned with the relationships between the in the last section of the chapter as a geobotanical form, composition and distribution of vegetation com- technique that uses remote sensing technology. This munities and their relationship to environmental factors, historical review of the place of geobotany in geology, particularly in undisturbed terrain (Cole, 1984). Since underscores the genesis and paradigmatic shift in vegetation type and characteristics can change with rock

Corresponding author: Beneah D. Odhiambo University of Venda, Private Bag X5050, Thohoyandou, 0950; Limpopo Province, South Africa. e-mail: [email protected] ISSN: 22536302 (print)/ISSN: 22536515 (on line) ©Editaefa DOI: 10.5616/ijgr 160004

28 B.D. Odhiambo type, vegetation has been used to derive geological infor- visual fashion (i.e., in ways that are not detectable by the mation and to discriminate parent rock material. This is human eye), a realisation that led to the development of because vegetation reflects the interplay of many envi- remote sensing applications in geobotanical studies. It is ronmental factors including the geochemistry of the soil noteworthy that the remote sensing concept was first at the rooting depth, the type of soil, the aspect, and even envisaged by Socrates in c.400 BC, when he observed variations of micro-relief. It is therefore useful for geolo- that “man must rise above the earth's surface, to the top gical mapping and mineral exploration. of the atmosphere, in order to understand the environ- Geobotanical prospecting utilizes the principle of ment in which he lives" (in Odhiambo, 1993). selective adaptation by plants to variations in geological When vegetation becomes stressed and develops a conditions in the subsurface. This includes observations distinct spectral characteristic in the infrared part of the of the presence or absence of certain plant species or electromagnetic spectrum, identification of “spectral vegetation assemblages, observations of changes in anomalies" at the red edge has enabled researchers to abundance and vigor around a dispersion halo, and identify mineralized areas using remote sensing identification of conspicuous morphological modifi- (Singhroy, 1989, and Odhiambo, 1993). cations in individual plants or in a group of plants (Dorr Historical Development of Geobotany et al., 1971). Thus, geobotany is based on the premise that plants, through their systems, absorb The use of plants in the search for subsurface mineral geochemically available elements from the subsurface resources has been in practice since ancient times. In and thereby reflect the underlying geological conditions Rome, during the reign of Augustus Caesar (63 BC to 14 (Siegel, 1974). AD) the architect Vitruvius stated that water is to be It is important to realize that the term geobotany is sought in areas where certain are to be found frequently used to include biogeochemistry. While bio- growing naturally and not artificially planted (Agricola, geochemistry is concerned with the detection of biogenic 1556). Physiological effects of metal on vegetation were haloes of ore indicator elements in living organisms and also observed by Agricola (1556), while Barba described their remains (Kovalevsky, 1987) using rapid and the use of plants for mineral prospecting in Potosi, sensitive analytical instruments (Brooks, 1972), in Bolivia in 1637 (Cannon, 1979). Other significant stu- practice it is difficult to treat geobotany and biogeo- dies that formed the fundamental basis for geobotanical chemistry as being mutually exclusive. This is because in methods of mineral prospecting include the works of virtually all geobotanical investigations it is common Lomonosov (1763) and Kapinsky (1941), both quoted in practice to sample and analyse the vegetal component for Malyuga (1964). In all these studies, emphasis was its constituent metallic elements when attempting to placed on observation of the depauperating effects of determine the “composition" of the subsurface geology. mineralization on whole plant communities. However, it In other words, to ascertain that a plant is indeed a was not until the early parts of the 19th Century that geobotanical indicator species, trace element analysis geobotanists propounded the notion of indicator species must be done in order to correlate the elemental for specific mineral types. Today, more descriptive terms concentration in the plant species with the substrate such as accumulators, specialists, or super tolerant plants geochemistry. Consequently, this trend of thought has are used more freely, caution being taken in declaring reached the point where most researchers consider and plants as indicators of specific mineral types (Cannon, treat geobotanical and biogeochemical methods of 1979). prospecting as constituting a single discipline. This is It is important to underscore what geobotany is appropriate since the absence or excess availability of a understood to mean among geological fraternity. Ques- certain mineral element may result in the anomalous tioning whether ecology, plant geography, and geobota- appearance and/or development of a plant, or even whole ny are three different sciences, or three different view- plant communities. Malyuga (1964) referred to the use of points of the same science and whether one is a part of plants for mineral exploration as the biogeochemical or the others, or if all three are synonymous, Rübel (1927) geobotanical method. The technique is used widely in the strived to correct the misunderstandings by non-ecolo- exploration for ore deposits, in the study of the character gists, by discussing what ecology includes in the minds and depth of groundwater, and even in the determination of ecologists, what plant geography means to plant geo- of the presence of geological structures, which are often graphers, and what geobotany comprehends. Rübel associated with ore deposits. In this chapter these terms (1927) recognized geobotany as that part of are used interchangeably. which has to do with gea, the earth, the action of all According to Raines and Canney (1980), the appli- earthly factors, all changes of plants on the earth, and all cations of geobotany can be summarized into three distributional questions over the earth. Essentially what conceptual approaches. First, is the study of plant com- Rübel stated is that geobotany (-plant geo- munities, including characteristic florae and specific graphy) is the science of the relationship of plants to the indicator species; second, is the study of vegetation environment, the earth. The article by Tadros (in Vege- density, which includes extreme cases of complete tatio, 1949) entitled, Geobotany in Egypt (A historical absence or presence of vegetation; and third, is the study review) does not attempt to define the term, but proceeds of . Each approach involves to use it from a botanical perspective. investigations of the various ways in which individual In the 3rd Edition of the Great Soviet Encyclopedia plants or whole plant communities respond to mineral (1970-1979; in Free Dictionary, 2010) geobotany is re- concentrations in the geochemical environment. It should ferred to as , a science concerned with be noted that the response can also affect plants in a non- the Earth’s vegetation as an aggregate of plant commu- The Place of Geobotany in Geology 29 nities, or phytocoenoses. Essentially, these positions Fortescue (1980) designed to illustrate the complexity stem from the statement of the German naturalist A. and the caution necessary when attempting to interpret Humboldt, which dates back to the beginning of the 19th the geology of an area using vegetation, particularly with Century, that vegetation is a unique element of nature. It the aid of remote sensing data with little or no ground should be pointed out that in these studies and academic truth and auxiliary data. It is evident from this figure that discourses the use of plants from a geological perspec- a given vegetation community can represent several tive as originally envisaged by architect Vitruvius (in geological conditions. Agricola, 1556) is not addressed. The term geobotany Thus, regional geobotanical remote sensing studies cannot be stretched to include palaeobotany; a view pro- have been met with varying degrees of success. This is pounded in the Great Soviet Encyclopaedia. because the success of the technique depends to a large Recent geobotanical studies have been undertaken in extent on plant response to the varied and complex several countries, especially in New Zealand and environmental factors, from the species level to the Australia (Brooks, 1972), in the Scandinavian countries community level. From the many case studies on (Rune, 1953; Talvite, 1979), USSR (Lomonosov, 1763; regional geobotanical remote sensing research carried Kapinsky, 1941), Canada (Fortescue, 1980; Wagner, et out in different parts of the world, some general trends al., 1989; Singhroy, 1989), and in the USA (Siegel, 1974, that can be grouped into three main categories have been 2013). In Africa, indicator ground-based geobotanical identified. work has been done in Botswana and Namibia (Cole, 1980; 1984; Cole et al., 1986), Zaire (Duvigneaud and Brenan, 1966), and in Zimbabwe (Wild, 1974; Brooks and Yang, 1984). In the East African region, similar studies have also been undertaken with respect to fluorite and chromite mineralization (Odhiambo, 1988; Odhiambo et al., 1989) and in Burundi (Singhroy, pers. comm., 1990). Chromite indicator plants were identified in Zimbabwe by Wild (1974). These plants aided greatly in the exploration and delineation of chromite deposits in that region. The levels at which plants are used to discriminate geological differences in the landscape include the use of plants for general geological mapping (regional geobotany) purposes, and for specific identification of mineral deposits (indicator or target geobotany). How mineral uptake influences the spectral reflectance charac- teristics of plants is an attribute of spectral geobotany. It should not be over emphasised that these entry levels are at the forefront of the developments in geobotanical applications. These entry levels are briefly discussed in the following sections. Regional Geobotany The term regional (or background) geobotany was introduced by Singhroy (1987) and adopted by Bruce Figure 1: Hypothetical vegetation zonation patterns in Alaska (After, Fortescue, 1980) and Hornsby (1987) in attempts to make the geobotanical concept more applicable in an exploration setting. It relies on the geobotanical concept at the community First is the use of vegetation for general geological level, rather than at the species level. As a result, it does mapping; this is based on the structural factors of not require the use of special sensors onboard an aircraft vegetation response. Second are analyses of the spectral and instead, satellite imagery can be used, thereby response characteristics of whole vegetation commu- reducing exploration costs in terms of data acquisition nities to excess availability (or absence) of certain mi- and processing. The regional concept is based on the neral elements in the subsurface - which leads to spectral premise that the nature and distribution of plant anomalies in the vegetation. Third are observations of the communities occur as a result of environmental taxonomic factors of vegetation response to different conditions and therefore variations in distribution from geological conditions. the established norm for an area may be indicative of The application of remote sensing data and different anomalous changes in the geology. The model, therefore, image enhancement techniques has been used by several requires an understanding of the general ecological researchers in attempts to identify suitable enhancements conditions of an area in terms of the relationships for the digital data that best depict distinct lithological between the bedrock units, the sacrificial materials units. (including soils, topography and plant communities) and In a study undertaken in the Spanish-Portuguese py- the resulting spectral patterns observed on remotely rite belt, Banninger (1985) made a comparison between sensed imagery. Landsat MSS and TM data for geobotanical prospecting. Figure 1 show five hypothetical case studies by In a similar study, Aranoff et al. (1986) used an integra-

30 B.D. Odhiambo ted approach in which a deposit model, Landsat imagery, of lithological units and the location of mineralized sites and sacrificial geochemistry were used to develop a in Africa and in the United Kingdom. In this study, procedure for locating tungsten mineralization associated imagery was also used in monitoring the seasonal with shallow buried intrusions. Although in these studies vegetation changes related to water stress in Botswana it was found that the mineralized sites with distinct and the UK. It was found that the detection of mineral geological units and mineralization could be delineated deposits on remotely sensed images depends on the using satellite data, it should be noted that in the identification of the geobotanical anomalies at the classifications used the pixels that were known to fall on community level. In Botswana, it was found that large mineralized areas were given distinct values during geobotanical anomalies that are confined to the ground image classification in order to locate the known sites of layer can be identified on satellite-based remotely sensed the mineral deposits. All the data were represented in the images if the canopy is sufficiently open and if form of co-registered images which constituted an image uptake of a toxic element by the trees causes contrasts in data base. reflectance with that of the neighbouring vegetation. Cole et al. (1986) used imagery of differing spectral They also concluded that the size of a deposit is an and spatial resolutions from Landsat MSS, Landsat TM, important factor for detection mineral anomalies in these SPOT and airborne Daedalus II line scanner for studies areas using remotely sensed data.

unclassified pixels correspond to shadowed areas.

Plate 2: Unsupervised classification Landsat TM in which Plate 1: FCC Landsat TM showing the structural geo- natural vegetation classes (Geobotanical Units) are morphology and geobotanical units West Pokot, Kenya identified

Bedell (1987) used edge-detection analysis of Indicator Geobotany satellite imagery for detecting gold deposits in Tanzania. Indicator geobotany recognizes the vegetation com- He found that remote sensing can assist the exploration position and/or communities that occur on mineralized geologist by differentiating structures both temporally areas as indications of associated mineral deposits. From and spatially when investigating the frequency and a historical perspective, it is emphasised that indicator geometry of edges in plan view. He found that this is geobotany has evolved to include elemental analysis of particularly important when use is made of different the mineral composition in plants that are associated with Gaussian filters and false colour composites of specific mineral deposits. In this regard, vast literature multitextural data, as they provide suitable images for exists with respect to indicator geobotany. Many works lineament interpretation. However, lithological units that have been published by the USSR (Russian) were not discriminated in the study. Academy of Sciences were extensively reviewed by Regional geobotanical techniques have also been Cannon (1960); Malyuga (1964); Chikishev (1965); used for traversing un-mapped geological terrain Brooks (1983); Kovalevsky (1987); Cwick (1987); (Singhroy, 1989; Bruce and Hornsby 1987; Bruce, pers. Odhiambo (1993). comm., June, 1992) during helicopter surveys in Guyana. The uptake of chromium and its influence on In this regard, the advancements in geobotanical vegetation is used in this section to demonstrate the techniques using remote sensing data for geological realm of indicator geobotany. Several researchers (Rune, applications lies in the identification of lithological units 1953; Proctor and Woodell, 1971; Morton, 1992; Lamb, which have been correlated to distinct vegetation 1993) have noted that serpentine rocks have distinct communities (Plate 1). Image classification can involve vegetation communities. Indeed, the presence of unu- application of standard statistically based decision rules sually sparse over ultramafic rocks has led to many in order to identify the land cover types for pixels in the studies of the reasons for this apparent infertility. Dating image. Unsupervised classification can be undertaken back as far as the 16th Century (in 1583), Caesalpino prior to the fieldwork with the objective being of (cited in Proctor and Woodell, 1971) described a plant identifying natural classes that can be recognised and restricted to the Upper Tiber Valley in Tuscany near described in the field. The result of the classification is Florence, Italy. Such vegetation is sometimes referred to presented in Plate 2. In this classification most of the as the serpentine barrens (Lamb, 1993). Serpentine soils The Place of Geobotany in Geology 31 have been shown to have very toxic effects on vegetation also be clearly identified on remotely sensed data. The (Malyuga, 1964; Lyon et al., 1968). The chromium species Protea kilimandischarica is a characteristic content of non-serpentinic soils is in the order of indicator of the area around the chromite mineralization. 100µg/g (ppm), whereas in serpentinic soils, chromium It is an evergreen and does not shed its . Satureja values around 5000µg/g (0.5%) are common; however, abyssinica is closely associated with the chromite de- the range varies from 1000 to 25,000µg/g, with the latter posits in the area; its tiny purplish-pink together commonly found in heavily leached tropical serpentine with its sweet minty smell, makes it a conspicuous herbal soils (Jaffre, cited in Brooks, 1987). As a result, these species. The Acacia - Dodonea - Combretum - Ficus ve- soils develop distinct vegetation communities. For getation community surrounding the chromite mine- example, Mouat (1982) observed that the foothills of the ralization is similar to that recorded by Brooks and Sierra Nevada are covered by an open Ceanothus Malaisse (1985) around the Great Dyke in Zimbabwe. chaparral with scattered digger pine (H. sabiniana), They also noted this vegetation community as not being while surrounding metasediments and metavolcanics are exclusive to the known chromium and nickeliferous covered by an oak woodland and grassland. Similarly soils. Miekel (in Mouat, 1982) described serpentine and non- Plate 3 shows a water lily which belongs to the plant serpentine vegetation in the Appalachian Piedmont. family Nymphaeaceae, growing in the mine water in Miekel noted that the serpentine is characterized by a abandoned Pb-Zn mine water was found to contain stunted tree flora with an open canopy dominated by concentration value of more than 10,000ppm lead Virginia pine (Pinus virginiana), post oak (Quercus (Odhiambo, 2014). Since water lily is an apparent stelleta), and blackjack oak (Q. Marilandica). Non- accumulator of lead, it can be effectively used in serpentine flora on the other hand is dominated by a bioremediation of the mine waters that have high lead robust tree flora with a closed canopy often dominated concentrations. The analytical data obtained from the by chestnut oak (Q. prinus L.) and white oak (Q. alba). study justifies detoxification of the mine waters. Proctor and Woodell (1971) stated that the most conspicuous feature of the Lizard Peninsula in Cornwall is the restriction of Minuartia verna to serpentine soils. In the serpentine endemic plants of the Great Dyke in Zimbabwe, Brooks and Yang (1984) reported a maxi- mum chromium concentration of 77µg/g (ppm). Similarly, Jaffre et al. (1979) reported a mean of 45µg/g in 17 species (132 specimens) of Geissois from New Caledonia, where Lee et al. (1977) found less than 10µg/g chromium in plants growing in soils containing up to 1% (10,000µg/g) of this element. Wild (1974a) also recorded high concentration of chromium (2400µg/g) in dried and upto 48,000ppm in ashed leaves of Sutera fodina, a Zimbabwean serpentine endemic. Wild (1974) also found 30,000ppm in the leaves of Plate 3: Water lily (family Nymphaeaceae) in Dicoma niccolifera, and up to 15,000ppm in the leaves Pb-Zn mine waters. of Pearsonia metallifera. Brooks (1987) believed that the very high concentration values of chromium in plant ash Spectral Geobotany could be due to contamination by wind-blown dust from This section elucidates the historical advancement of the Noro chrome mine, since other specimens of the the remote sensing concept that was first envisaged by same species from other parts of the Great Dyke gave a 1 Socrates in c.400 BC. Spectral geobotanical approach in maximum value of only 2µg/g chromium in dried leaves geobotanical remote sensing has also been referred to as (Brooks and Yang, 1984). It should be noted that the target geobotany (Singhroy, 1987). It crystallizes the geochemical conditions (mainly pH and Eh values) may relationships between spectral properties of plants and be completely different in the area where Wild's (1974) the geochemical stress phenomenon. There is a need samples were collected, when compared with the sites therefore to establish an understanding that the where Brooks and Yang (1984) sampled. These geochemical condition of the substrate influences environmental conditions are not stated in either study. spectral reflectance properties of plants (Odhiambo, How the samples were prepared (i.e., whether ashed, or 1993). just dried and then digested) is also not indicated. Today, spectral geobotanical investigations are Geobotanical results, presented by Odhiambo (1993) widely used in geobotany since leaves of healthy, reveal a clear decrease in species diversity from non- actively growing plants produce broadly similar spectral mineralized Precambrian Basement gneisses to mine- reflectance curves (Horler et al., 1980; Goetz et al., ralized serpentine host rocks. Variation in species com- 1983; Milton and Mouat, 1984; Singhroy, 1989). position is exemplified by the characteristic Protea - Spectroradiometers with narrow spectral band working Faurea - Maerua - Satureja vegetation community that in the vegetation reflectance edge (the near infrared occurs on the mineralized geobotanical unit, which can portion, also called the red-edge) of the electromagnetic spectrum have been developed in order to measure and 1 Researches tend to express results differently (e.g. ppm or quantify the distinct spectral changes. The variations in µg/g). Samples are also prepared differently.

32 B.D. Odhiambo absorption and spectral reflectance characteristics of should be established in a direction perpendicular to the plants have been related to pigment content, cellular general strike of the lithological units. The data used in structure, and to the moisture content of leaves (Cwick, geobotanical sampling research should be collected 1987). Odhiambo (1993; 2008) used the SE590 during two field seasons that should be selected to Spectroradiometer to analyse spectral reflectance coincide with the driest time (months) of the year when characteristics (from 380nm to 1200nm) of vegetation in vegetation is most stressed, and mineral concentrations relation to mineral concentrations. in the subsurface are at their highest (Cole, 1983; Lyon et Hyperspectral (imaging spectroscopy) remote sensing al., 1982). In both cases the sampling period should take is currently being investigated by researchers with regard as short a time as possible in order to minimise climatic to the detection and identification of minerals, terrestrial variations that might influence element uptake by plants. vegetation, and man-made materials. The concept of Sampling should be during the same time of day to avoid hyperspectral remote sensing began in the mid-80 and is any diurnal climatic impacts on the vegetation. currently being used widely by geologists for the Wet seasons should be avoided since leaching of mapping of minerals. In hyperspectral imaging, actual elements from plant leaves may occur (Brooks, 1983), detection of materials is dependent on the spectral and concentrations in the subsurface may be diluted by coverage, spectral resolution, signal-to-noise ratio of the the excess water available in the soil due to heavy spectrometer, on the abundance of the material and the rainfall. A second reason for the dry season being ideal strength of absorption features for that material in the for geobotanical studies is that the atmosphere is rela- wavelength region measured (Lee and Landgrebe, 1993). tively cloud-free. As a result, conditions for collection of The degree of association between chromite spectral data are ideal and cloud free satellite imagery is pathfinder elements and independent spectral parameters easily selectable (Odhiambo, 1993). Finally, fieldwork is at the red edge of the electromagnetic spectrum raises the not impeded by impassable roads, which can occur question as to whether or not a purely prospective (as during wet seasons in most tropical environments. opposed to a retrospective) approach to mineral exploration using geobotanical methods is feasible. The Geobotanical Samples results presented by Odhiambo (1993) indicate that it is possible both at the general level of interpretation and at Geobotanical samples should be separated (where applicable) into twigs and leaves; the reason being that the species specific level. In either case, a positive correlation implies that as the concentration of an the two plant parts (organs) are known to concentrate element increases in the sample, the value of the different amounts of trace elements. In addition, each sample should be split several times to provide independent spectral parameter also increases. Conversely, a negative correlation implies a decrease in replicates. This enables repeated determinations to test the value of the independent spectral parameter as the the repeatability of analytical results obtained. Standard concentration of the element increases. plant material should be used (or prepared) with which to compare the results from the analyses. General Methodology for Geobotanical Studies Sample preparation in geobotanical trace element analysis is of utmost importance, since, for example, the The historical aspects of the methodologies used in ashing technique preconcentrates the elements in plant geobotanical studies for mineral exploration are not material. The concentration levels so obtained are clearly elucidated in the literature surveyed. The usually much higher (sometimes by a factor of about 20 methodology is to be found only in the contemporary fold) than the values obtained for the same samples when studies summarized in this section. the material has simply been dried and then digested In planning a geobotanical study, the sampling (Odhiambo, 1993). procedures adopted should take into account the Several studies (Brooks et al., 1985; Odhiambo, 1988; potential differences in metal uptake by different plant and Odhiambo, 2015) have shown that strong corre- species and plant parts. In addition, the stages of lations exist between soil and plant element conce- maturity and seasonal cycles should be considered, as ntrations. So, when preparing for fieldwork and data recommended by Thornton (1986). This is of practical collection, one should endeavour to find out if any importance because in temperate countries the maximum previous mineral evaluation studies around the study "elemental peak" in plants is restricted to the first week area had been done. If the results of such a mineral of spring (Canney et al., 1979; Banninger, 1985). In assessment program are available and the quality of the these regions, this factor greatly limits the timing of results are acceptable for the purposes of the research, fieldwork for geobotanical studies. The timing of the then no additional geochemical analysis of the soils and studies also has a bearing on the spectral discrimination rock samples is necessary. It is recommended that only a of annual and perennial plants in both temperate and few soil samples may be collected from selected sample tropical climates. During the dry season in the tropics, sites (from soil sampling pits) in order to determine the annual plants are absent and spectral reflectance from soil pH and Eh, soil organic matter content and soil perennial plants becomes more conspicuous on satellite texture, since these physicochemical factors control images. uptake mineral elements by plants. Although Malyuga (1964) and Brooks (1972) have Normally, a brief reconnaissance survey where recommended that two or three of the more dominant transects and sampling plots are established on the plant species be sampled so that at least one of them is geobotanical units map should be undertaken in order to present at each sample point, experience using a pre- identify locations for collecting samples. Transects The Place of Geobotany in Geology 33 selected sampling grid has shown (Malyuga, 1964; elements that are associated with the chromite Kapinsky, 1941 cited in Brooks, 1972; Brooks, 1984; mineralization (the pathfinder elements, namely nickel, Cole, 1986; Kovalevsky, 1987; Odhiambo et al., 1989; cobalt, and chromium) are non-essential plant micro- Odhiambo and Howarth, 1993a) that geobotanical nutrients. These heavy trace elements are not known to indicator plants are not necessarily the dominant species. be essential for normal development of vascular plants Thus, all species should be sampled indiscriminately (Siegel, 1974; Raven et al., 1976 cited in Singhroy, from each sample point, during the first field season. 1989). The non-essentiality in is an im- This biogeochemical approach to mineral prospecting portant factor in geobotanical studies since it has also using vegetation is based on the continuum concept been shown that the uptake of essential elements (such as (Elliot, 1983), where vegetation can be used as calcium, potassium, sodium, phosphorous, and sulphur) biogeochemical samples without the need for mapping is controlled by plants (Kovalevsky, 1987). In trace vegetation units. Furthermore, with respect to elemental element analysis this factor (non-essentiality) usually re- uptake by plants, it has been demonstrated (Kovalevsky, sults in negative biogeochemical anomalies (Odhiambo, 1987) that some plants are more informative (non- 1988). Availability of the trace elements to plants has barrier) while others are not (barrier type), as illustrated been observed to result in physiological and/or morphological changes in plants due to their toxicity (Cannon, 1960; Malyuga, 1964; Brooks, 1972; Rose et al., 1979). In plant samples, statistical analyses of the trace element concentrations show typical negative skewness which is characteristic of geochemical data (Siegel, 1974; Al Ajely, 1984). The data can therefore be subjected to standard statistical treatment, in order to establish whether or not the concentration values obtained are true geochemical anomalies within the geochemical environment. Using all the concentration values obtained for each element assayed in a study, statistical threshold levels are calculated to determine the mean element concentration, local background concentration value, and local threshold concentration value for all the elements. The local background value Figure 2: Types of biogeochemical samples (LBV) is determined as the mean element concentration plus one standard deviation, while the local threshold in Figure 2. value (LTV) is the mean concentration plus two standard At every pre-selected sample point, plants should be deviations. Concentration values higher than the local sampled after a careful scrutiny of all the plants at that threshold value are considered to be anomalous, as location in order to observe any visible signs of recommended originally by Marmo (1958), Sayala morphological anomalies. Each plant sample should (1979), and Rose et al. (1979). Once these thresholds are consist of the apical leaves of the plant. In the case of ascertained, concentration plots can be constructed using tree species, leaves are collected from several locations the biogeochemical data (Odhiambo, 1993). Results from (limbs or branches) around the canopy. These are then a study around the chromite deposits in Kenya showed mixed to make one sample of approximately 10 to 15g in that smoothing biogeochemical data rids it of some weight. This recommendation is based on the result of spurious concentration values. Further, known sites of several studies which have shown that the elemental chromite mineralization are clearly depicted and the content of leaves can vary greatly from one side of a serpentinite host rock is also clearly depicted by the plant to another, depending on the location of the mineral chromium, nickel and manganese anomalies. deposit. There seems to be a longitudinal passage of The synthesis of the element concentration plots inorganic material from the of one side of a plant to performed by Odhiambo (1993) using the standard the limbs on the same side (Cannon, 1964; Singhroy, smoothing technique shows that, even though the pers. comm., Feb., 1993). technique rids the data of some spurious concentration During the second field season, geobotanical values, it is still fairly accurate in locating the known sampling and spectral measurements should be sites of the chromite/nickel deposits, especially when undertaken simultaneously. Plants found in proximity to results are compared to those obtained by the a mineralized site during reconnaissance and are also in experimental plots. the non-mineralized sites should be sampled. In this case It is concluded that concentrations of the chromite sampling can then be focused on the two or three pathfinder elements in geobotanical samples causes dominant species (as recommended by Malyuga, 1964, measurable spectral shifts at the red spectral edge of and Brooks, 1972) that were found by analyses from first associated vegetation. The spectral response of plants to season data to respond well to the mineralization. element concentration is species specific since, a given In the chromite mineralization studies undertaken in plant species responds uniquely to mineralization when West Pokot District of Kenya, Odhiambo (1993) found compared to another plant species. This observation that with the exception of manganese, the other mineral concurs with observations made in the works of Masuo-

34 B.D. Odhiambo ka (1981), Labovitz et al., (1983) and Singhroy (1989). The first recommendation concerns the lack of Secondly, the spectral response of plants is element standard reference material for use in geobotanical specific. In this case, the effects of one element on the studies. There are two issues; first is the lack of any independent spectral parameters were observed to be adequate reference standards. Secondly there are also no unique from those of another element. The chromite reference standard samples against which the spectra of mineralization shifts the red edge of the associated plants different plants and their derived independent spectral towards shorter wavelengths. reflectance parameters can be compared. The second Summary issue necessitates the identification of an appropriate plant as a reference standard. The importance of selecting pathfinder elements for The second recommendation stems from the analysis, together with the need to use high resolution availability of chromium (and other elements) in the spectroradiometers for analysis of the spectral reflec- biogeochemical environment. It is evident that the tance characteristics of plants for mineral concentration chromite-tolerant plant species are associated with the in the subsurface has been underscored (Odhiambo, mineralization. Therefore, the general lack of awareness 2008). The chromite indicator species were identified on among exploration geologists and environmentalists, as the basis of high accumulations of the pathfinder to the power of geobotanical research in mineral elements, coupled with their persistent occurrence in exploration, and the niche it occupies in environmental proximity to the chromite mineral deposits. Results from biogeochemistry and geoepidemiology raises the need to West Pokot District of Kenya case study demonstrate carry out further studies. The effectiveness of the how chromite deposits can influence the associated technique in mineral exploration should be assessed, in vegetation. exploration programs. The following conclusions are made from the study The third recommendation concerns measurement of undertaken in West Pokot District of Kenya. First is the pH and conductivity (Eh) conditions of associated soils. concern of lack of standard reference material in These conditions are not usually presented in the results geobotanical studies. These are essential to ensure the of similar studies. They are essential in understanding quality of the biogeochemical data. The second geochemical availability of an element around a mineral conclusion concerns the use of remote sensing since deposit. These parameters should always be measured. mineralized geobotanical unit are unique with respect to Similarly, sample preparation technique used should their spectral reflectance characteristics. It is evident that always be clearly presented. This is because, for the delineation of geobotanical units is of particular example, ashing of plant material preconcentrates the importance for geobotanical remote sensing studies. elements present in the sample as compared to plant The anomalous concentrations of chromium were material that is simply digested and then analysed. Such found in Satureja abyssinica, Leucas tomentosa, and in an approach would facilitate comparisons of the Protea kilimandischarica. In these plants, the amount of concentration ranges of chromium (or other elements of chromium does not exceed 500 ppm. These concen- interest) in similar geobotanical studies from different trations are much lower than the chromium concentration regions. values obtained by Wild (1974) from the serpentine Lastly, spectral geobotanical datasets are usually very endemics around the Noro Chromite Mine area in Zimbabwe. However, they are similar to the values large and should therefore be archived in a "central obtained by Brooks (1972; 1987) and Brooks and Yang spectral data bank". This would facilitate quick (1984). Therefore, for purposes of prospective geobo- referencing of the spectral reflectance parameters of tanical mineral exploration, the need to identify non- plant samples from different geographical regions of the barrier plants like Protea kilimandischarica, Faurea world. saligna and Satureja abyssinica seems critical. References It is concluded that concentrations of the chromite pathfinder elements in geobotanical samples causes Agricola, G., 1556: De Re Metallica, (Translated from the first measurable spectral shifts at the red spectral edge of Latin Edition of 1556) Dover Publications Inc., New York, associated vegetation. The spectral response of plants to 1950, 79p. element concentration is species specific since, a given Al Ajely, K.O., Andrews, M.J., and Fuge, R., 1984: Biogeo- plant species responds uniquely to mineralization when chemical dispersion patterns associated with mineralized porphyrystyle mineralization in the Coed y Brenin forest, compared to another plant species. This observation North Wales; in Prospecting in areas of glaciated terrain, The concurs with observations made in the works of Institute of Mining and Metallugy, pp.1-10. Masuoka (1981), Labovitz et al., (1983) and Singhroy Aranoff, S., Goodfellow, W., Bonham-Carter, G.F., and (1989). Secondly, the spectral response of plants is Ellswood, D.J., 1986: Integration of surficial geochemistry element specific. In this case, the effects of one element and Landsat imagery to discover skarn tungsten deposits on the independent spectral parameters were observed to using image analysis techniques; Proc. of IGARSS’ Symp., be unique from those of another element. The chromite Zurich, Switzerland, ESA SP 254, pp.513-520. mineralization shifts the red edge of the associated plants Banninger, C., 1985: Comparison between Landsat MSS and towards shorter wavelengths. Thematic Mapper data for geobotanical prospecting in the Following from the conclusions presented above, five Spanish - Portuguese pyrite belt, International Geoscience and recommendations made below arise mainly from the Remote Sensing Symposium; Proc. of IGARSS’ Symp., studies undertaken in Kenya and from observations made Amherst, M.A, v.2, pp.949-956. over the years in publications on geobotanical appli- Brooks, R.R., 1972: Geobotany and Biogeochemistry in cations. Mineral Exploration, Harper and Row, Publishers, 290p. The Place of Geobotany in Geology 35

Brooks, R.R., 1987: Serpentine and its vegetation: a multi- Fortescue, J.A.C., and Singhroy, V.H., 1985: A metalanguage disciplinary approach; Ecology, Phytogeography and for spectral geobotany; 19th Intenational Symposium on Physiology Series Vol.1, Dudley, T.D (Ed), Dioscorides Press, Remote Sensing, pp.849-854. Portland, Oregon, 454p. Gauthier, R.P., and Neville, R.A., 1985: Narrow-band mul- Brooks, R.R., and McDonnell, M.J., 1983: Delineation of New tispectral imagery of the vegetation red reflectance edge for Zealand ultramafic rocks by computer processing of digital use in geobotanical remote sensing; Proc. 3rd Int. Colloquium data from satellite imagery; New Zealand Journal of Science, on Remote Sensing on Spectral Signatures of Objects in v.26, pp.65-71. Remote Sensing, Les Arcs, France, ESA SP-247, pp.233-236. Brooks, R.R and Yang, X.H., 1984: Elemental levels and Goetz, A.F.H., Rock, B.N., and Rowan, L.C., 1983: Remote relationships in the endemic serpentine flora of the Great sensing for exploration: an overview; Economic Geology, Dyke, Zimbabwe and their significance as controlling factors v.78 (4), pp.523-590. for the flora; Taxon, v.33(3), pp.392-399. Horler, D.N.H., Barker, J., Darch, J.P., , D.C., and Ba- Bruce, B., and Hornsby, J.K., 1987: A Canadian perspective on rringer, A.R., 1983: Approaches to detection of geochemical the application of satellite remote sensing to regional stress in vegetation; Adv. Space Res., v.3 (2), pp.175-179. geobotany; Geocarto International, v.3, pp.53-59. Hornsby, J.K., Bruce, W.D., Harris, J., and Rencz, A.N., 1988: Bruce, B., June, 1992: On the history and evolutionary deve- Implementation of background and target geobotanical lopment of geobotany and spectral geobotanical techniques techniques in mineral exploration; Them Conf Remote for geological mapping and mineral exploration purposes, Sensing for Exploration Geology, VI, 11p. Toronto, Ontario, Canada; Personal Communications. Howard-Williams, C., 1971(a): Environmental factors con- Canney, F.C., Cannon, H.L., Cathrall, J.B., and Robinson, K., trolling the growth of plants on heavy metal soils; KIRKIA, 1979: Autum colors, insects, plant diseases and prospecting; v.8, pp.91-102 Economic Geology, v.74, pp.1673-1692. http://www.csr.utexas.edu/projects/rs/hrs/process.html Cannon, H.L., 1979: Advances in botanical methods of pros- Jaffre, T., Brooks, R.R., and Trow, J.M., 1979: Hyperaccumu- pecting for minerals, Part I - Advances in Geobotanical lation of Nickle by Geissois species; Plant Soil, v.51, pp.157- Methods; in: P.J. Hood (Ed.), Geophysics and Geochemistry 162. in the Search for Metallic Ores; Geological Survey of Canada, Economic Report 31, pp.385-397. Kettig, R. and D. Landgrebe, 1976: "Classification of mul- tispectral image data by extraction and classification of Chikishev, A.G., (Ed), 1965: Plant Indicators of Soils, Rocks, homogeneous objects", IEEE Transactions of Geoscience and Subsurface Waters; Consultants Bureau Enterprises Inc., Electronics, Vol. GE-14, No. 1, pp. 19-26, January 1976. New York, New York, 210p. Kovalevsky, A.L., 1987: Biogeochemical Exploration for Cole, M.M., 1980: Geobotanical expression of ore bodies; Inst. Mineral Deposits 2nd Ed., Trans. from the Russian, Brooks, Mining and Metallurgy, (London), pp.B73-B91. R.R (Ed), VNU Science Press, Utrecht, The Netherlands, Cole, M.M., 1984: Integrated use of remote sensing imagery 224p. and geobotany in mineral exploration; Trans. Geological Labovitz, M.L., Masuoka, E.J., Bell, R., Siegrist, A.W., and Society of South Africa, v. 56, pp.13-28. Nelson, R.F., 1983: The applications of remote sensing to Cole, M.M., Farago, M.E., Mehra, A., and Vujakovic, P., 1986: geobotanical exploration for metal sulfides - results from the Remote sensing imagery for the detection of stress in 1980 field season at Mineral, Virginia; Economic Geology, vegetation caused by drought and/or metal toxicity; 20th Int. v.78, pp.750-760. Symp. Remote Sensing of Environment, Nairobi, Kenya, Lamb, L., April 1993: On The serpentine barrens of North pp.889-898. America; Waterloo, Ontario, Canada, Personal Communi- Cwick, G.J., 1987: A Geobotanical Assessment of the Silver cations. Mine Area of Southeastern Missouri using Landsat TM Data; Lee, C. and D.A. Landgrebe, 1993: "Feature extraction based PhD Dissertation, Indiana State University, Terre Haute, on decision boundaries", IEEE Trans. on Geoscience and Indiana, USA, 146p. remote Sensing, 15(4): 388-400, 1993. Daubenmire, R. F., 1966: Vegetation identification of typal Lee, G.W., Brooks, R.R., Boswell, R.D., and Jaffre, T., 1977: communities; Science, v.151, pp.291-298. Plant-soil relations in New Caledonia serpentine flora; Plant Dorr, J. Van H., Hoover, D.B., Offield, T.M., and Shacklet It, Soil, v.46, pp.675-680. H.T., 1971: The application of geochemical, botanical, geo- Lefevre, M.J., 1980: Teledetection et anomalies geocliniques a physical and remote sensing mineral prospecting techniques Echassiere (Allier, France); Chronique Rescherche Maniere, to tropical area; USGS Open File Report. v.453, pp.41-43. Duvigneaud, P., and Brenan, J.P.M., 1966: The genus Lyon, G.L., Brooks, R.R., Peterson, P.J., and Butter, G.W., Cryptosepalum Benth. (Leguminosae) in the areas of the 1968: Trace elements in a New Zealand serpentine flora; Flora of Tropical East Africa and the Zambeziaca; Kew Bull., Plant Soil, v.29, pp.225-240. v.20, pp.1-23. Lyon, R.J.P., Elvidge, C., and Lyon, J.G., 1982: Practical re- Elliot, J.C., 1983: Vegetation inventories in the Fort Union quirements for operational use of geobotany and biogeoche- Region; in: Coal Development Collected Papers Volume II; mistry in mineral exploration; Proc. Int. Symp. Remote Papers Presented at the Coal Development Workshops in Sensing of Env., 2nd Thematic Conference: ’Remote Sensing Grand Junction, Colorado and Casper, Wyoming, USA, for Exploration Geology’, Fort Worth, Texas, v.1, pp.85-96. pp.1061-1086. Malyuga, D.P., 1964: Biogeochemical Methods of Mineral Free Dictionary, 2010: 3rd Edition of the Great Soviet Prospecting, Consultants Bureau, New York, 205p. Encyclopaedia Milton, N.M., and Mouat, D.A., 1984: The significance of scale Fortescue, J.A.C., 1970: A research approach to the use of in geobotanical applications for lithological discrimination vegetation for location of mineral deposits in Canada; Taxon, and mineral exploration; in: Remote Sensing for Geological v.19, pp.695-704. Mapping, Teleki, P., and Weber, C., Eds., Proc. of the seminar Fortescue, J.A.C., 1980: Environmental geochemistry: a habitat of IUGS - UNESCO Programme on Geological Applications approach, Ecological Studies 35, Springer-Verlag, New York, of Remote Sensing (GARS), pp.287-298. 347p.

36 B.D. Odhiambo

Morton, J.R., 1990: On the general concept of indicator plants Richie, J.C., 1962: A geobotanical survey of northern in Africa, and the occurrence of salt indicator species around Manitoba; Arctic Inst. of North America, Technical Paper, Windsor, Ontario, personal communications. No.9, 48p. Mouat, D.A., Arp, G.K., Collins, W., Elvidge, C., Labovitz, Rose, A.W., Hawkes, H.E., and Webb, J.S., 1979: Geochemis- M.L., Lyon, R.J.P., Milton, N.M., Parrish, J., Rock, B.N., and try in Mineral Exploration, 2nd Edition; Academic Press, Wickland, D.E., 1983: The importance of geobotany in London, 657p. geological remote sensing applications; in: Hederson , F.B., Rübel Eduard, 1927: Ecology, Plant Geography, and and Rock, B.N., (Eds.); Frontiers for Geological Remote Geobotany; Their History and Aim. Botanical Gazette, v.84 Sensing from Space; Report of the 4th Geosat Workshop, (4) 1927, pp.428-439. Flagstaff, Arizona; American Soc. Photogramm., v.43, pp.57- URL: http://www.jstor.org/stable/2470620 64. Rune, O., 1953: Plant life on serpentines and related rocks in Odhiambo B.D.O., 1993: Influence of Chromite Mineralization on the north of Sweden; Acta Phytogeogr. Suec, v.31, pp.1-139. Vegetation: a Geobotanical Investigation in West Pokot District, Siegel, F.R., 1974: Applied Geochemistry; John Wiley and Kenya; PhD Thesis, University of Waterloo, Ontario, Canada, Sons Inc. London, 353p. 248p. Singhroy, V.H., 1989: Geobotanical Remote Sensing in The Odhiambo B.D.O., 2008: Spectral Reflectance of plants as Canadian Boreal Forest Region, PhD Thesis, State University indications for Mapping Economic Podiform Chromite of New York, College of Environmental Science and , Mineral Deposits in west Pokot District, Kenya; East African Syracuse, New York, USA, 393p. Journal of Applied Science, Vol.1 (1) 2008. ISSN 20700903. Singhroy, V.H., 1987: An overview of geobotanical remote Odhiambo B.D.O., 2006: Geobotanical Indicators for Fluorite sensing in Canada; 11th Canadian Symp. on Remote Sensing, Mineralization in the Savannah Woodlands of Kerio Valley, University of Waterloo, Waterloo, Ontario, pp.259-269. Kenya in JEANARM, Vol. 2(1), 2006. ISSN 1684-0755. Tadros, T.M., 1949: Geobotany in Egypt (A historical Review). Odhiambo B.D.O. and P. J. Howarth., 2002: Chromium Vegetatio, v.2 (1) 1949, pp.38-42. Concentrations in Geobotanical Samples from West Pokot URL:http://www.jstor.org/stable/2003448 District, Kenya: Journal of Environmental Geochemistry and Health, v.24, pp.111-122, 2002. ISSN 0269-4042. Talvitie, J., 1979: Remote sensing and geobotanical prospecting in Finland; Bull. Geol. Soc. Finland, v.51, pp.63-73. Odhiambo B.D.O., 2002: The Use of Plants in Mapping hazardous concentrations of toxic elements in Mineralized Tomlinson, R.F., Brown, W.G.E., 1962: The use of vegetation Areas; Africa Geoscience Review, Vol. 9, No.4, pp.377-384, analysis in photointerpretation of surface material; 2002. 1117-370X/2002 [Printed in France]. Photogramm. Eng. Remote Sensing, v.28 (4), pp.584-592. Odhiambo, B.D.O., 2014: Mapping and assessing Wagner, C.L., Howarth, P.J., and Singhroy, V.H., 1989: A environmental and health impacts around Vitengeni lead-zinc spectral geobotanical study at Natal Lake, Northern Ontario; abandoned mine in Coastal Kenya. In: Kríbeč, B. and Davies, Int. Journal of Remote Sensing; v.10 (11), pp.1721-1734. T.C. (Eds.): Addressing Environmental and Health Impacts of Wei Wen-Zhong., 1978: Genesis and genetic types of chromite Active and Abandoned Mines in Sub-Saharan Africa, p.145– deposits in Kansu, China; in: Chemical Abstracts, v.90, 150; Czech Geological Survey; ISBN: 978-80-7075-858-8. pp.58246j. Prodgers, R.A., 1984: Collection and analysis of baseline Wild, H., 1965: The flora of the great dyke of southern vegetation data, Minerals and the Environment, v.6 (3), Rhodesia (Zimbabwe) with special reference to the serpentine pp.101-104. soils; KIRKIA, v.5, pp.49-86. Raines, G.L., and Canney, F.C., 1980: Vegetation and geology; Wild, H., 1974: Indigenous plants and chromium in Rhodesia; in: Siegal, B.S., and Gillespie, A.R., (Eds.); Remote Sensing KIRKIA, v. 9, pp.233-242. in Geology, John Wiley and Sons, New York, pp.365-380. Zhumar’, A. Yu., 1993: Effects of vegetation-cover parameters Rediske, J.H., 1956: Chromium toxicity in plants; in: Biology on the location of the maximum slope of the ’red edge’; Research Annual Report; Richland, Washington, Hanford Soviet J. Remote Sensing, v.10 (4), pp.594-603. c Atomic Products Operations (Report HW-41500), pp.48-52.