The Research Of Characteristic Of Geochemical Spatial Distribution In Ambatondrazaka, Madagascar

The author of this thesis had been in Madagascar as a technician of Geological Survey Team of Madagascar Assisted by China from June to December of 2006, and engaged in the item of geochemical survey in Ambatondrazaka, scale of which is 1:50000. Based on this program, the characteristics of geochemical spatial distribution of this district were mostly investigated in large scale. Due to the Non-disclosure Agreement and weakness of the basic geological work of Madagascar, the author just takes the data of a mapsheeet (1:25000) for example, expecting to find a math way of seeking anomaly of spatial combination, locating the anomaly of element association on this studied district and trying to speculate its characteristics of geochemical spatial distribution.To analysis the characteristics of geochemical spatial distribution of a district, people mainly use the math way to process the geochemical exploration data, find the anomaly, then explain it on the ground of its actual geological situation. Traditionally, geochemical exploration anomaly is valued by single element, in other words, it is determined by the extent of the content of the simplex element. The multi-element problem is out of consideration or just plus the result of single element mechanically. The process of forming mine is pluralistic, which means that geochemical elements of forming mine is pluralistic, the phase of forming mine is multistage and the condition of forming mine is various. So it is reasonable to infer that geochemical exploration anomaly which reflects mine's anomaly should be multielement composite anomaly, single element's anomaly had ripped the composite relation, and factor analysis can better reflect this composite relation.According to the national criterion, we had done lots of outdoors actual survey in large scale, and got the first-hand samples. After doing geochemical analysis, 38 kinds of elements had been found in those samples. 19 of them were chosen in this article, and were processed by R-factor analysis to get the principal factor solutions, based on which the orthogonal solutions with varimax rotation were figured out, including the factor loads and factor scores.The multi-element variables of the original data matrix were divided into some indicative element associations by factor loads. After that, the author processed every R-factor score of those element associations in statistical way, ruled out the high score points step by step, worked out the average background value( X ) and standard deviation(δ) of each factor, then determined the threshold( T1 ) and upper limit( T2 ) of every element association's anomaly, drew the map of element association's anomaly, explained it in geological way, and induced the characteristics of geochemical spatial distribution in this studied region.19 kinds of elements'content of 459 gridded specimens in the studied region were processed by R-factor analysis to get the orthogonal factor loads'matrix with varimax rotation. Afterwards, six types of element associations were summed up identified as follows: the basic-ultrabasic element association, the high or medium temperature's one, the oxyphile one, the low temperature's and thiophile one, the chromium one, and the acidic one. After doing statistic analysis to the orthogonal R-factor scores with varimax rotation of six element associations, the high score points( X + 3δ)were eliminated gradually till the residuary data were at the interval(X -3δ,X + 3δ).And then, the average background value, standard deviation, threshold and upper limit of anomaly were worked out.Based on the element association's anomaly profile made by above-mentioned parameters, the characteristics of geochemical spatial distribution in this studied region were outlined as follows: The first element association's entire anomalies are positive anomaly and 14 places are located, including 3 places of biggish area. All of the second kind ones are positive anomaly and 19 places are located, including 4 places of biggish area. All of the third kind ones are positive anomaly and 10 places are located, including 3 places of biggish area. Not only positive anomalies but also negative anomalies are located about the fourth kind ones, nevertheless their area are baby-size. There are not only positive anomalies (11 places, including 4 places of biggish area) but also negative anomalies (5 places, including 3 places of biggish area) about the fifth kind ones, and 16 places are located. There are not only positive anomalies (15 places, including 4 places of biggish area) but also negative anomalies (5 places, including places of biggish area named HS-10) about the sixth kind ones, and 20 places are located.The novelty of this thesis is the employment of factor analysis to get better element association, and then determined the threshold and upper limit of anomaly by common way, found the anomaly of each element group, finally induced the characteristics of geochemical spatial distribution. This method makes up for the defects brought by the traditional way (the method of single-element or mechanism stacking technique of multi-element).Generally speaking, this method is available and efficient, but there is still great space for improvement.