The karst process acts as carbon sequestration for atmospheric CO2 . The amount of karst carbon sequestration (KCS) depends on the discharge of karst catchment and inorganic carbon concentration of the water body. Based on the data from the monitoring station on Banzhai subterranean stream located in Maolan National Nature Reserve of Guizhou province, the process and influence factors of KCS have been analyzed. It shows that the amount of KCS is about 353 t C per year in the catchment of Banzhai subterranean stream, and there is good linear relationship between the strength of KCS and discharge of the stream at various time scales. Therefore, how to monitor the discharge accurately is the key to the estimation of KCS. And stations with real-time monitoring function are very important for KCS calculation because of strong seasonal variability of the karst water cycle.
karst carbon sequestration ; monitoring ; Banzhai subterranean stream ; Discharge ; Rainfall ; HCO3−
In the studies of global carbon cycle, known carbon dioxide sources and sinks are imbalanced. There is an unknown carbon sink, which has caused widespread concern among scientists. Studies have shown that there is a Northern Hemisphere atmospheric carbon dioxide sink [ Fang et al ., 2001 ; Siegenthaler and Sarmiento, 1993 ] and its strength accounts for about one third of the missing global carbon sink [ Wang et al., 2002 ]. This may be related to the fact that 67% of the global land is in the Northern Hemisphere, but the mechanism is not clear and there are many uncertainties [ Fang et al., 2001 ]. Therefore, the geological carbon absorbing effect has gradually attracted the attention of scientists [ Kump et al., 1994 ; Hartmann et al ., 2009 ; Gaillardet et al ., 1999 ; Probst et al ., 1994 ].
For a long time, it was generally accepted that the geological carbon sink rate was very small, thus its effect was not considered seriously in the global carbon cycle study. But, recent research has found that geological carbon sink is also actively participating in the carbon cycle, and its magnitude has been underestimated in the previous researches. Karstification, participating in carbon cycle, holds the most important share in the geological carbon sink. According to Blum et al.  , Jacobson et al.  , and Gao et al.  , even if the exposed area of the carbonate rock strata is very small, most of the HCO3− flux in the river is from the karstification.
China is a big country with vast karst. The carbonate rock distribution is widespread, with an area of 3.44 million km2 , one third of the total land area of the country, or 1/7 of the world exposed area of carbonate rocks. Because of the formation conditions and the particularity of the geographical distribution, karstification is very strong [ Yuan, 1993 ; Yuan and Zhang, 2002 ], the karst water cycle caused carbon absorbing effect is extremely significant [ Yuan, 1993 ; Jiang et al ., 2000 ; Yuan, 2001 ; Xu and Jiang, 1997 ]. Since the 1990s, Chinese scientists have devoted to the karst research. During 1990–2010, four international geological comparison projects (IGCP299, IGCP379, IGCP448, and IGCP513) led by Chinese scholars have transformed China’s geographical advantages in karst studies into academic advantages. Especially project IGCP379 — the karst processes and the carbon cycle, received high remarks from the international peers [ Goldscheider et al., 2008 ]. However, in the karst area in southwestern China, the seasonal variability of water cycle is very large, the temporal and spatial variation of karst carbon sequestration (KCS) is also rather large. In order to study the role of geological carbon sequestration more precisely, quantifying KCS seems necessary to construct a geological carbon monitoring network. In the background of international carbon emission negotiations and under the support of the achievements of preliminary projects, the China Geological Survey started the project of “Potential Evolution of Geological Carbon Sink in China” aiming at constructing a monitoring network to improve the estimating accuracy of geological carbon sink.
This paper takes Guizhou Banzhai underground river monitoring station located in Maolan National Nature Reserve as an example to analyze the karst carbon process and influence factors. It is the authors’ hope that this research will provide reference for the layout of the subsequent national geological carbon monitoring network sites.
The Banzhai underground river is located at the center of the Maolan National Nature Reserve, which is part of Libo county, Guizhou province. On geomorphic units it is located in the slanted zone transitioning from the south Guizhou Plateau to Guangxi hills plain [ Zhou , 1987 ], with an elevation of 450–1,100 m and 80% coverage of primary forest. To study the karst water cycle carbon absorbing effect of slope zone covered by primary vegetation, in October 2009 automated monitoring stations were built on the surface outlet river of the underground river system, which covered an area of 32.25 km2 . Monitoring stations have a mulriple weir of triangle and rectangular. At the monitoring station CDTP300 multi-functional hydrochemical online monitoring instrument made by Greenspan Company, Australia, was installed. The testing items include pH value, electrical conductivity (EC ), water level, water temperature, and precipitation; and their precisions were 0.01, 10 μ s cm–1 , 0.01 m, 0.1°C, and 0.5 mm, respectively. The time interval for measurements was set as 15 min.
Karstification carbon is mainly inorganic carbon which can be approximately represented by the concentration of HCO3− . The reaction equation is:
where Me denotes the metal ion in carbonate rocks, in general it is Ca or Mg.
Due to the limitation of the present instrument, the concentration of HCO3− cannot be monitored online. However, it has a good linear relationship with electrical conductivity. In this research area, fitting of EC and HCO3− concentration resulted in the relationship between the two as follows:
where [HCO3− ] denotes HCO3− concentration (g L–1 ).
Discharge can be calculated with the water table and parameters of the weir through the empirical formula [ Fetter, 2001 ]. It can be seen in Eq. 1 , that in the karst carbon sink half of HCO3− ions are from the atmosphere or soil CO2 , whereas the other half are from the carbonate rocks. Thus, the above-mentioned method can be used to obtain the discharge of the underground river system and the HCO3− concentration, and the karst carbon amount can then be calculated as follows [ Liu et al., 2010 ]:
where Kcs is karst carbon amount (kg C); Q is discharge (m3 per month); Δt is computing time interval (month); 12 and 61 are molar mass for C and HCO3− respectively. Coefficient of 0.5 indicates half of the HCO3− in karst water is from the atmosphere or soil CO2 .
For nearly a hydrological year from November 1, 2009 to October 20, 2010, hydrological and hydrochemical data were obtained. Through the compiling and analysis of real-time monitoring data, rainfall, discharges, karst carbon amount, and the runoff modulus and runoff coefficient were calculated. The results were shown in Table 1 .
|Time (year-month)||Precipitation (mm)||Discharge (103 m3 )||Carbon sin (t C)||Runoff modulus (mm)||Runoff coefficient (%)|
Table 1 shows that the maximum runoff coefficient appears in May 2010, with a value of 52.5%, but the precipitation is only 225.5 mm. Monthly precipitations for both June and July are more than 350 mm, but runoff coefficients are only 37.7% and 32.0%, respectively, which are even less than that of January when the total rainfall is merely 73 mm. Comparison of the precipitation patterns shows that precipitations in May scatter around with small amount during each rainfall event and short periods of no rain; whereas in June and July precipitations are relatively concentrated with great amount during each rainfall event and longer period of no rain. Epikarst zones covered by forest have a water deficit during the former no rain period; the later precipitations must overcome the early deficit to produce runoff [ Ran et al ., 2002 ; Jiang et al ., 2008 ]. Therefore, precipitation pattern and the effect of the underlying surface make rainfall and discharge positively correlated, but it is not a strictly linear correlation.
According to Eq. 3 , karst carbon sink is influenced by HCO3− concentration and discharge, and both should be positively correlated. In this study, we analyzed the dynamic change of the intensity of carbon sink on the storm-scales of four precipitations from July 10, 2010 to August 19, 2010. On the storm scale, the variations of HCO3− concentration in the water is very large. And macroscopically, the carbon sink intensity of this underground river catchment and water body HCO3− concentration are negatively correlated (Fig. 1 ). This shows the relationship between HCO-concentration and the intensity of carbon sink is rather complicated, which involves different reaction mechanisms [ Liu et al., 1999 ]. Correspondingly, the dynamic variations of the intensity of carbon sink and the discharge are almost in synchronization, showing good linear correlation.
The relationship curve between karst carbon sink (KCS ) values and discharges, and HCO3− concentration at storm time scale
On the monthly time scale, karst carbon sink amount and discharge still show good linear relationship (Fig. 2 ). Rainfall in May is obviously less than that in July; contrarily the karst carbon sink amount is significantly greater than that in July. This indicates karst carbon sink amount is not controlled by precipitation. Underground water discharge on the basin is the main controlling factor of the karst carbon sink.
The relationship curve of monthly KCS and discharges
The aforementioned phenomenon shows that, although the underground river discharge change leads to a big change in HCO3− concentration, the magnitude of the former change is far greater than the latter, thus the contribution of HCO3− concentration change is covered by that of the discharge change. In this process, the magnitudes of both variations are great, strong spatial-temporal differences are manifested. Any unreasonable estimates would result in significant errors in carbon sink estimation. Therefore, real-time on-line monitoring is really necessary.
(1) The intensity of karst carbon sink is influenced by the river basin discharge and the concentration of inorganic carbon; macroscopically it is mainly controlled by the discharge of the river basin.
(2) Seasonal variability of karst carbon sink is very big, as a result, the use of high resolution real-time monitoring technology is essential to accurately estimate and evaluate the intensity and potential of karst carbon sink.
The authors would like to thank the Maolan National Natural Reserve for allowing access and providing help with our research. This work was funded by the project (No. 41072192) from National Natural Science Foundation of China and the project (No. 1212011087122) from China Geological Survey.