Effect Of Different Fertilization On CO₂ Emission For A Paddy-upland Rotation Purple Soil Under Wheat

As an important carbon pool, agricultural field system exchanges carbon with atmospheric carbon pool through soil respiration, which directly affected global carbon cycle. However, the factors effecting soil respiration are various and corresponding mechanisms are still complex. For understanding the effect of long-term fertilization on flux of carbon dioxide (CO2) and dissolved organic carbon (DOC) in purple soil at wheat growing period, this study based on eight different fertilization levels treatments in National Purple Soil Fertility Monitoring Station, at Beibei of Chongqing, China, from November 2008 to May 2009, investigated changes of CO2 flux among different treatments by close chamer method: FnMR-(NPK normal fertilization+manure+no crop residue); FnR- (NPK normal fertilization + no crop residue); R-(no fertilization+ no crop residue); R+(no fertilization+crop residue); PKR-(PK fertilization+no crop residue); NR- (N fertilization+no crop residue); FnR+ (NPK normal fertilization+crop residue); FhR+ (NPK high fertilization+crop residue)。The detailed targets in this study were:1) The effect of long-term fertilization on the content and change of DOC; 2) The effect of long-term fertilization on CO2 emission; 3) Revealing the key factor controlling the emission of CO2. The main research results were listed as follows:Ⅰ. Effect of long-term fertilization on DOC dynamics and other parameters for paddy-upland rotation purple soil under wheatThe DOC average annual values of different treatments ranged from 56.62～117.7 mg/kg, but the differences were not significant.The DOC annual dynamic changes of all treatments from November to December in 2008 kept in lower levels (8.12 mg·kg-1～22.5 mg·kg-1), beginning with February 2009, DOC of soils increased significantly with a great obviously fluctuation (10.32～532.09 mg·kg-1), from March to May 2009, DOC mainly concentrated on low levels(12.53～92.75 mg·kg-1).During the wheat season treatments without N-fertilizer application including R+, R- and PKR-, all kept lower levels without obvious fluctuation. The treatments with N-fertilizer application, corresponding index showed an initial increasing then following a decreasing process. The TDN of FnMR-, FnR-, NR- and FnR+ all reached peak values in late December, and the peak of FnMR- was 421.07 mg·kg-1, which was more than twice higher the other treatments. The average values of TDN in R+,FnR+,FhR+ were 10.18 mg·kg-1,25.55 mg·kg-1 and 55.91 mg·kg-1 respectively, which all had significantly positive correlation with nitrogen application through linear fitting.Ⅱ. Effect of long-term fertilization on soil microbial biomass carbon (MBC) and microbial biomass nitrogen (MBN)The soil MBC of different treatments ranged from 95～216 mg·kg-1. The average value of MBC in PKR- was 95 mg·kg-1,612%,56.0% lower than NR- and R-,the difference was obvious. The MBC of FnR- and NR- were 1.37 and 1.26 times higher than PKR- and R-,but the effect was not obvious (p>0.05).Contrasts the MBC in the sites with same crop residue condition, R+>FnR+≥FhR+; NR->R->FnR-, but the differences were not obvious. Among the sites with same fertilization, the MBC in R+ was 44.4% higher than R-, and the MBC in FnR+was 30.9% lower than FnR-.MBN of R+,FnR+,FhR+ averagely were 73 mg·kg-1,72 mg·kg-1 and 65 mg·kg-1 respectively, but which showed no obvious differences (p>0.05). Similarly, the differences among NR-, R- and FnR- were not obvious. However, the treatment FnMR-, which is only treatment applied by fertilizer, MBN was 85 mg·kg-1, which significantly was higher than the others and difference was obvious(p<0.05).Ⅲ. Effect of long-term fertilization on CO2 emissionFor the average value of CO2 flux, R+ was 25% higher than R-, but the difference was not obvious. FnR+was 48% higher than FnR-, the difference was obvious(p<0.05). Contrast the there treatment with crop residue, CO2 flux showed R+≈FnR+>FhR+, and for the treatment without crop residue, R->FnR-≈NR-, the differences were obvious(p<0.05).From December in 2008 to January in 2009, soil temperature decreased and reached bottom in late January, meanwhile, the CO2 flux decreased. The CO2 flux from all sites except R+ and FnMR- reached bottom in late January also. Then the CO2 flux increased with the increasing soil temperature.The seasonal variation of cumulative CO2 emission from different sites from December 2008 to January 2009 increased slowly with insignificant differences (9.57kg·ha-1～910.38 kg·ha-1). From January to May 2009, the cumulative CO2 emission increased fast, and there exist gaps among different sites.Ⅳ. Factors controlling CO2 emission Stepwise regression analyses showed that soil temperature could explain 59.1% of the temporal variation in the instantaneous soil CO2 flux at FnR+, the soil temperature together with WFPS could explain 74.9% of CO2 flux at site R-, FnR- and PKR- at which no relationship was found between soil CO2 flux and soil properties. At site FnR- and PKR-, soil DON explained 23.7% and 26.5% of soil CO2 flux. DON also has a certain influence at sites FhR+ and NR-.For understanding the relationship between CO2 average flux and physical-chemical characteristics of soils, the CO2 average flux and soil factors derived from different treatments were all linear fitted. The results showed, CO2 average instantaneous flux had positive correlation with pH of soils (R2=0.5754, p=0.048)and MBC(R2=0.7015, p=0.019)respectively, and negative correlation with NO3- (R2=0.5211, p=0.043).The cumulative CO2 emission of the there treatment with crop residue showed R+>FnR+ >FhR+, which had significantly negative correlation with nitrogen application through linear fitting(R2=0.995, p<0.001).Long-term fertilization significantly affected soil physical, chemical and biological parameters such as pH, NO3- and MBC, etc. then had complex regulatory effects on soil respiration. The most essential is that single or high application of N had significant inhibitory effect on soil respiration.