Relationship between carbon dioxide/methane emissions and the water quality/sediment characteristics of Taiwan's main rivers

River and sediment have unique carbon dynamics and are important sources of the dominant greenhouse gases (GHG), carbon dioxide (CO2) and methane (CH4). To understand the relationship between CO2/CH4 emissions and water quality/sediment characteristics, we have investigated critical parameters in the river water. Eight parameters of water quality (dissolved oxygen, oxidation-reduction potential ORP], chemical oxygen demand, biochemical oxygen demand BOD5], suspended solid, nitrate NO3-], NH4+, and bacteria) and four sediment characteristics (total organic carbon TOC], total nitrogen T-N], NO3-, and ammonium NH4+]) were measured in two of the larger rivers in Taiwan, and relevant environmental conditions were recorded. The experimental results indicated that CO2 emissions from the river were mainly affected by BOD5 concentrations and the levels of bacteria. CH4 emissions, on the other hand, were greatly affected by the ORP in the river. The correlation between CO2 emissions and sediment characteristics was insignificant (R2 < 0.3). However, TOC and T-N in the sediment may lead to increases in CH4 emissions into the atmosphere. A deeper analysis of the relationship between the different parameters and GHG emissions by ANOVA and the multiple regression method revealed that CO2 emission (y) was significantly related to bacteria number (x1) and BOD concentration (X2). The regression equation takes the form y = 0.00032x1 + 3.18089x2 + 25.37304. Also, the regression relationship between CH4 emission (y) and ORP (x) in the river can be described as y = -0.825216x + 169.02257. The relationship between CH4 emission and sediment characteristics may be described as y = 5.073962x1(TOC) + 2.871245x2(T-N) - 12.3262. Extra sampling data were collected to examine the feasibility of the developed multiple regression equations. The experimental results suggest that the emissions of such GHGs as CO2 and CH4 from rivers can be predicted using the regression equations developed here. Moreover, the emissions may be reduced by manipulating the proper factors.