Rain and Lake Waters in taiwan:Composition and Acidity

Ten automatic samplers were used to collect dry deposits and rain water in the highly industrialized region of southwestern Taiwan between December 1985 and July 1989. During the dry season, rain water sometimes appeared basic because it contained a large amount of basic minerals such as calcium carbonate released by several local cement plants. During the wet season, however, the rain became increasingly acidic with its lowest pH value of 3.823. Water samples collected from 137 lakes and reservoirs in Taiwan and its six largest offshore islands were analyzed. According to the rock and soil compositions, the present authors divided Taiwan into three regions. Zone I consists mainly of igneous rocks with lakes generally of low buffering capac­ ity. The major catio.ns are Na+(0.280 ± 0.335rneq/1; 50.3%), Mg+2(0.120 ± 0.168meq/1; 21.5%), ca+2(0.112 ± 0.094meq/1; 20.0%), K+(0.032 ± 0.022 meq/1; 5.7%) and H+(0.014 ± 0.015meq/l; 2�5%). T he major anions are Cr � (0.269 ± 0.329meq/l; 46.0%), HC03(0.204 ± 0.438; 34.8%), S0,!2(0.104 ± 0.043meq/l; 17.7%), N03(0.007 ± 0.008meq/I; 1.3%), P0,!3(0.001 ± 0.003meq/1; 0.18%) and OH-(0.0001 ± 0.0002meq/I; 0.02% ). Zone II consists mainly of non-calcareous sedimentary and metamor­ phic rocks With lakes of medium buffering capacity. T he major cations are ca+2(0.729 ± 0.801meq/1; 52.2%), Mg+2(0.437 ± 0.464meq/I; 31.3%), Na+ (0.203 ± 0.226meq/1; 14.5%), K+(0.019 ± 0.017meq/l;1.4%) andNHt(0.008 ± 0.015meq/l; 0.6%). The major anions are HC03(0.888 ± 0.933meq/1; 52.2%), S0,!2(0.527 ± 0.687meq/1; 31.0%), C2 HsC00-(0.212 ± 0.507meq/l; 12.5%), Cl-(0.056 ± 0.083meq/1; 3.3%), N03(0.013 ± 0.019meq/1; 0.8%) and P043(0.003 ± 0.008meq/l; 0.2%). Zone III consists mostly of gravel, sand, clay, limestone and the alluvium zone with lakes of high buffering ca­ pacity. T he major cations are ca+2(0.909 ± 0.953meq/1; 41.6%), Na+(0.630 1 Institute of Marine Geology, National Sun Yat-Sen University, Kaohsiung, Taiwan, · R.O.C.


INTRODUCTION
Lately acid rain has attracted much attention and concern worldwide. In Taiwan, the adverse effects of acid rain, the offspring of rapid industrialization, have also been felt and discussed. Lakes are showing both direct and indirect responses to the acidification of water by acid rain. Directly, acid water affects the growth of organisms in lakes, changes the ecosystem (Scheuharnmer, 1991), affects the growth of crops if used for imgation, and speeds up the corrosion of generators and metal pipes if used to generate electricity (Likens and Bormann, 1974;Likens et al., 1979;Glass et al., 1982;Kramer and Tessier, 1982). More circuitously, the release of heavy metals by acidification originally adsorbed on sediments results in high concentrations of heavy metals in the lake biota. Similarly, the heavy metal concentrations in crops also increases if acidified water is used for irrigation. These processes pose serious health problems regardless of whether the acidifi ed lakes constitute the source of drinking water, or whether man consumes food which has been irrigated with acid water and has accumulated high concentrations of heavy metals (Morel and Morgan, 1972;Scheider et al., 1979;Davis et al., 1982;Kilham, 1982;Campbell and Stokes, 1985).
The process of lake acidification that leads to the adverse effects, such as those men tioned above, obviously should be investigated thoroughly. Unfortunately, because of the lack of historical data, the writers do not know clearly whether the lakes in Taiwan have changed or will soon change in terms of their acidity (Chen et al., 1992). The acidity of rain and lake waters in Taiwan and the sensitivity of these lakes to acid rain based on pH and alkalinity data are reported here. These data are hereby used to reveal spatial variations and could be used as a baseline for future temporal comparisons. Additionally, possible future changes in lake acidity based on studies of total dissolved salt, rock and soil types (Henriksen, 1979) are also estimated _ in this paper.

MATERIAL AND METHODS
Ten automatic dry/wet samplers were used to collect dry deposits and rain water in southwestern Taiwan between December 1985 and July 1989 ( Figure 1). These dry deposits were collected periodically and examined by a scanning electron microscope (JEOL JSM-35 CFSEM) and by X-ray diffraction (Diano 8536 XRD). Over two hundred wet samples were collected and analyzed within 24 hours after the rain. The measurements made included pH, alkalinity, conductivity, density, S04 2 , Cl-, No;, ca + 2 , Mg +2 , K + , Na + and NHt .
NIST(National Institute of Standards and Technology, USA) pH 4.004 and 7.415 buffers were used to calibrate the Radiometer PHM 85 system, precise to ± 0.003 units. Alkalinity was

1191150' 120�00·
Ki n� m n v24,30, measured by the Gran Titration (Chen, 1988) and density by a vibrating tube densitometer (Chen et al., 1980). A Dionex 2000i Ion Chromatography was used to measure K + , Na+, NHt , ca + 2, Mg+2, p-, Cl-, No;, Br-and S042, while a Perkin Elmer 2380 AA was also used to measure ca+2, Mg+2, K + and Na+. Nutrients were measured by the Flow Injection Analysis. Certifi ed ERA (Environmental Resource Associate) rain water samples (2694I and 2694II) were used as references. The relative error is less than 6% Hung and Chen, 1989). A Millipore-double distillation system was used to provide pure water for reagent preparation and for blank detection. Over six hundred lake water samples were collected from 137 lakes ( Figure 1) either with a bucket (for small ponds) or with a Nansen bottle when boats were available. In remote areas without access to a boat, the authors used a remote-controlled vessel, which could be controlled to obtain samples from as deep as 15 m. Normally, samples were collected at three depths at two locations in each lake with temperatures and pH being measured immediately. Afterwards, once samples were filtered with a 0.45µ m Nalgene filter, they were stored in amber plastic bottles at 4 ° C, and shipped back to the laboratory for further analyses of the above mentioned parameters. The filter paper was pre-rinsed with the sample before being used for filtration. The parameters measured were the same as those for rain water. Reliability of the data was further checked by correlating the sum of the cations with the sum of the anions. A linear correlation was found (r=0.96, n=507) with a slope of 0.96, thus differing only slightly from the expected value of 1 ( Figure 2).
A composite rain water sample consisting of all of the samples collected between May, 1986 and June, 1988 was pre-concentrated 20 times by evaporation at sub-boiling tem perature. This sample was used to measure minor and trace elements by the Inductively Coupled Plasma (ICP) at National Tsin Hwa University (Hsu and Chen, 1993). Similarly, a composite of lake water samples collected between September, 1985 to July, 1987 was pre concentrated 20 times before measurements werl? taken by the ICP. Such pre-concentration  · was necessary because of limitations in the detectability of the ICP. Detailed sampling and analytical procedures were described elsewhere Hung and Chen, 1989).

Rain Waters
Dry deposits consist mostly of quartz, gypsum and aragonite probably coming from several cement plants in the sampling area. A one to fifty weight dilution of dry deposits with deionized/distilled water (pH=5.66) results in solutions with pH values ranging from 7.34 to 10.23 .
The pH range and the numerical means of rain waters collected each month are also shown .in Figure 3. The pH of rain waters collected covers a wide range due to the large differences in local conditions. pH values are sometimes quite high in the dry season, perhaps because the samples contain a relatively. large amount of airborne mineral particles.
A pH value as high as 10.635 is recorded, reflecting high aragonite contents. In general, the higher the alkalinity or conductivity, the higher the pH values appear. Additionally, the rain water becomes relatively more acidic during the wet season (May-Sept.), and there is some indication that it is less acidic at the onset of a rain. This phenomenon also suggests that basic particles contribute to part of the rain water samples. Later during a precipitation event, the water becomes more acidic because basic particles have previously been removed from the atmosphere .
Rain brought in by three typhoons in 1986 are less acidic. The mean pH values for normal rains in June through September in 1986 are slightly more acidic than the mean pH values for all rains combined ( Figure 3). The reverse is true� however, in 1987 and 1988.
Of the areas studied, Hsiaogun (site 9) seems to be the most aff ected by the acid rain because it is adjacent to the heavy industrial area concentrated in southern Kaohsiung. Rain waters there are frequently below pH 5. The amount of nitrate is inversely correlated with pH ( Figure 4); thus, nitrate apparently is a heavy contributor of acid rain at Hsiaogun. The pH also correlates inversely with excess sulfate (the amount of sulfate in rain water after correction for sea salts) (Figure 4), suggesting that excess sulfate also contributes to the formation of acid rain at this site. The effect of sulfate is potentially large because of its higher concentration than nitrate.
Based on limited data, the study indicates rain waters collected outside the Kaohsiung area generally have a pH value greater than 5.0. The sulfate and nitrate concentrations are also lower in the less industrialized region. These results reflect the relation between acid rain and air pollution.
In Table 1, the average rain water composition of 10 stations in Taiwan from 1990 to 1994 (Jeng, 1993)     As a first approximation, Taiwan is also divide into three major zones in this study according to the geologic compositions and geographic factors. Zone I includes the Tatun and Keelong volcano groups, the five offshore islands in northern Taiwan and Lutao and Lanyu. They consist mainly of igneous rocks such as andesite and andesitic pyroclas tics. These rocks have low buffering capacity, and the lakes in this region should have   Jeng, 1993) No.  0.003meq/l; 0.18%) and OH-(0.0001 ± 0.0002meq/I; 0.02%) are the major anions ( Table  2). The average salt content is l.169meq/l, while the average alkalinity is 0.204meq/l. Fur ther acidification of some of these lakes is possible because of the low buffering capacity of those lake waters and the surrounding rocks and sediments. and NHt (0.008 ± 0.015meq/l; 0.6% ). The major anions are HC03 (0.888 ± 0.933meq/l; 52.2%), so4 2 (0.527 ± 0.687meq/1; 31.0%), C2H5coo-co.212 ± 0.507meq/I; 12.5%),
Of particular interest is the Nan Jen Lake group in Kenting National Park. These lakes did not exist prior to 1980, but the waters contain mainly NaCl which probably comes from rain water and sea spray. It is not known to what extent this group of lakes is in equilibrium (or in a steady state) with their bottom sediments or the surrounding rocks and soils (acidic laterite). A more intensive study of these lakes is currently in progress.
Bicarbonate seems to contribute most to the alkalinity, and calcium concentration correlates well with alkalinity (r=0.89, n=48 l; Figure 5). Because of limitations in the detectability of ICP detection, a large sample had to be used in this study in order to pre-concentrate it for minor and trace element analysis by the ICP. As a result, composite samples were necessary leading to the possibility of bias in the data. Concentrations of certain trace metals in twenty-six events of rain water in 1990 which were pre-concentrated and measured with graphite furnace atomic absorption spectrophotom etry (Perkin-Elmer 5100PC HGA-600), however, show spatial and temporal variations in southwestern Taiwan (Lin and Hung, 1992). With the exception of Al, metal concentrations of composite rain water (1986)(1987)(1988), nevertheless, are covered by the range of samples in 1990. Since both the ICP and the atomic absorption methods give similar results, the data may be considered at least semi-quantitative. It is understood that these data (Table 3) are the only ones available in Taiwan, and a long-term observation may still be necessary to see the typical trace-element composition of rain water in Taiwan. Besides, the authors are particularly interested in V which may be a proxy for paleoclimate .
The concentration is much lower in lake water than in rain water, perhaps reflecting active biological removal in a lake. Since most concentrations are below 1 ppb, more detailed work is needed to show how reliable or how representative these values are.

CONCLUSIONS
Ten stations for automatic dry/wet precipitation samples were established in southwest ern Taiwan between December 1985 and July 1989. The chemical analysis of rainwater includes pH, conductivity, alkalinity, density, S04 2 , Cl-, N03 , ca +2 , Mg +2 , K+, Na + and NHt . The lowest pH value of the event rainwaters occurring in the stations of the Greater Kaohsiung area was 3.823, which was much lower than those found in the less industrialized neighborhood stations, where _the pH values of event rain waters were generally greater than 5.0. The concentrations of nitrate and sulfate of rain waters in the Greater Kaohsiung area were also higher than those found in the neighborhood stations, showing a strong correlation between rain water acidification and air pollution.
Samples were also collected from 137 lakes all over Taiwan. The results show that lake alkalinity is low in regions of igneous rocks but high for the sedimentary zone in western Taiwan, and variable for the eastern region with metamorphic rocks. Several alpine lakes without local pollution are acid-sensitive ones. In general, low-elevation lakes are high in pH and alkalinity and are related to the soil and rock types or man-made eutrophication from the inputs of fertilizers, pesticides and other agricultural chemicals. These lakes are not in any danger of acidification in the near future. Concentrations of 41 minor and trace elements in composite rain and lake samples are reported.