Method Of Real-Time Measurement In Ammonia Emissions From Farmland Based On Open-Path TDLAS Technique

Ammonia (NH3) is the most abundant alkaline gas in the atmosphere. It has many negative effects on ecosystem function and health, and on air quality. The volatilization of NH3 following fertilizer application is the important source of atmospheric NH3, and is an important path of nitrogen movement to the atmosphere, especially in calcareous soil. The main problem of study ammonia emission from farmland is to develop monitoring technique and method. Currently, the most techniques collect the atmospheric ammonia by an acid absorbent which often takes several hours to collect sufficient NH3 for analysis in the laboratory, too long for determining diurnal variation of ammonia flux, especially in cases of low concentrations. Subsequent laboratory analysis after sampling also requires a high labor input. The open-path tunable diode laser absorption spectroscopy (TDLAS) is a reliable and convenient tool for nonintrusive, continuous, real-time, online monitoring of ammonia concentrations in the lower atmosphere under the long optical path, and tackles the problem of rapid ammonia concentration measurement. The TDLAS technique is well suited for instantaneous monitoring of NH3 concentrations under field conditions. In this study, the TDLAS technique was integrated with backward Lagrangian stochastic (BLS) model to construct the high temporal resolution, high sensitivity and on-line measurement method for monitoring ammonia emission from farmland. The study assessed the applicability and capability of the TDLAS-BLS method for estimating NH3 emissions from farmland.The main results were summarized as follows:(1) The synthetic source experiment indicated that the environmental factors that influenced the accuracy of the TDLAS-BLS method were atmospheric stability(1/L) and friction velocity (u*). The BLS model underestimated emission rate (QBLS) under unstable condition(-300 m< L<-10 m) and overestimated emission rate under stable condition (10 m< L< 59 m). The most inaccurate emission rates were obtained under extremely unstable condition (0m<L<10 m) or extremely stable condition (-10m<L<0 m), while the most accurate emission rates were obtained under slightly unstable-near neutral conditions (L<-300 m或L>59m). The relatively great errors in QBLS were primarily associated with|L|< 5 m, which should be removed. The emission rate was also influenced by the friction velocity (u*). The uncertainty in QBLS with u*≤0.15 m·s-1 was large. Excluding data with u* ≤0.15 m·s-1 or|L|<5 m increased the recovery rate of the released gas (QBLS/Q) from 0.90 (σQ/Q=0.30, n= 117) to 0.97 (σQ/Q= 0.25, n= 90). The technical factors that influenced the accuracy of the TDLAS-BLS method were laser path height and fetch. The optimum measurement height was influenced by the fetch, atmospheric stability and friction velocity. The optimum measurement height was between 1.0 and 1.5 m at fetch of 15-30 m. QBLS/Q was insensitive to vegetation height when the height of laser path above plant canopy exceeded 0.4 m. The average QBLS/Q at different fetches were not significantly different at fetch of 15-60 m, indicating that NH3 did not reduce dramatically due to surface deposition or chemical transformation within 60 m. QBLS/Q at different fetches were mainly affected by the friction velocity and atmospheric stability. The NH3 concentrations in the center of the plume decreased by 36 percent as the fetch increased by one time. The laser sensor should be placed as close to the source as possible to reduce concentration measurement uncertainty. The averaging time of data used in the BLS model had little effect on the accuracy of QBLS.There were not significant differences between QBLS inferred from the different averaging intervals, even during times of rapid transition. The better choice is to use short averaging time to improve the temporal resolution of the modeled emissions.(2) The validation experiment indicated that the rejection criterion of |L|< 5 m was insufficient.|L|<10 m was sufficient to discard error-prone data for the longer range study. After eliminating data with u*≤0.15 m·s-1 or|L|< 10 m, the NH3 emission rates were not significantly different from those measured by the micrometeorological mass balance (MMB) method. On equivalent time scales, the TDLAS-BLS method estimated the total NH3 loss to be only 2.3% less than the MMB results. The validation experiment also indicated that the accuracy of the TDLAS-BLS method was insensitive to the averaging time. The TDLAS-BLS method also showed high accuracy in monitoring ammonia emissions from fertilizer applied to farmland. In the seedling stage of summer maize, the ammonia emission rates estimated by the TDLAS-BLS method were not significantly different from those measured by the MMB method. Cumulative ammonia emissions determined by the TDLAS-BLS method and the MMB method were similar, at 39.1 kg N·ha-1 (TDLAS-BLS) and 36.6 kg N·ha-1 (MMB), respectively. The ammonia emission rates from fertilizer applied to winter wheat estimated by the TDLAS-BLS method were 17% higher than those of the MMB method. There were also not significant differences between the total ammonia losses inferred from the two methods.The ammonia emissions measured by the static chamber (SC) method, which has been widely used to estimate ammonia emission from farmland in China, differed considerably from those estimated by the TDLAS-BLS method. Following application of urea to summer maize, there were significant differences between ammonia emission rates, patterns of ammonia emission, and total ammonia losses inferred from the two methods. At the topdressing stage of winter wheat, the patterns of ammonia emission measured by the two methods were similar due to long sampling time (≥24 h). The ammonia emission rates and total ammonia loss following surface application of urea to winter wheat measured by the SC method were significantly lower than those estimated by the TDLAS-BLS method; while after urea deep applied to winter wheat, there were not significant differences between the results estimated by the two methods. The reason was that relative low emission rates masked the differences between the two methods. Besides, the deep application of urea suppressed the effect of weather condition on ammonia emissions.(3) The combination of the TDLAS technique and the BLS model achieved continuous, real-time monitoring of ammonia emissions under field conditions, and obtained dynamic of ammonia emission via high-temporal resolution data. The results indicated that there were large diurnal variability and daily variability in ammonia emissions from cropland in the Huang-Huai-Hai Plain. Due to capricious weather in spring, the diurnal patterns of ammonia emission at the topdressing stage of winter wheat were various. There was small variability in the diurnal patterns of ammonia emission between different days at the topdressing stage of summer maize, since the diurnal variation of ammonia flux was closely tied to the solar radiation and temperature. For winter wheat, ammonia emission began at a low rate and increased slowly, presumably due to low temperatures. In the high temperature season, the ammonia emissions increased rapidly after urea applied to summer maize, and the following day reached emission peak of 26.7 kg N-ha-1·d-1 at seedling stage and 30.3 kg N-ha-1·d-1 at shooting stage. Cumulative ammonia losses for surface and deep application of urea to winter wheat were 18.5 kg N·ha-1 (13.3% of applied N) and 10.7 kg N·ha-1 (7.7% of applied N), respectively. The results indicate that the deep application of urea at the topdressing stage of winter wheat could substantially reduce ammonia emission compared with surface application. Cumulative ammonia loss after fertilization at seedling stage of summer maize was 39.1 kg N·ha-1 (22.4% of applied N), which was slightly lower than that at shooting stage of summer maize (46.4 kg N·ha-1,26.7% of applied N). Considering that the total ammonia loss at shooting stage did not include ammonia emissions at nighttime, the difference between the total ammonia losses of two growth stages. This was mainly because fertilization before irrigation at seedling stage reduced ammonia emissions, but fertilization after rainfall at shooting stage increased ammonia emissions.(4) The high temporal resolution measurements of ammonia emissions provided an opportunity to statistically analyze the factors that influence NH3 emissions. Ammonia emissions were influenced by environmental conditions, soil characteristics and management practices in the Huang-Huai-Hai Plain. At the topdressing stage of winter wheat, ammonia emissions after surface application of urea were mainly affected by wind speed and soil temperature, followed by solar radiation; while ammonia emissions after deep application of urea were closely related to solar radiation, followed by wind speed. This indicates that fertilizer application method appeared to dominate the environmental effects on ammonia emissions. Ammonia emissions at seedling stage of summer maize were significantly correlated with solar radiation and soil temperature. At shooting stage of summer maize, a dominant solar radiation relationship was found, and soil temperature effects were relatively weak. At the topdressing stage of winter wheat, ammonia emissions were not significantly affected by soil NH4+-N content, probably because high soil moisture content suppressed ammonia emission. At two growth stages of summer maize, ammonia emissions were all significantly correlated with soil NH4+-N content. This indicates high soil moisture content did not suppress ammonia emission, presumably due to high potential for NH3 diffusion in the high temperature season. Benefitting from the fine temporal resolution of the TDLAS technique, this research found the significant influences of high wind speed and rainfall on ammonia emission, indicating that ammonia emissions were mainly affected by the dominant meteorological factor in different weather conditions.