No abstract

No abstract

Abstract. Time domain reflectrometry (TDR) isFor calibration, undisturbed soil blocks, with a TDR probe installed in the center, were saturated and then allowed to dry by evaporation. Volumetric water content was calculated from measured K a values and from gravimetric measurements. Because we used undisturbed soil samples, our calibration accounts for the natural heterogeneity in soils. We tested the suitability of various calibration functions relating K a to soil water content for our soils. TDR technique underestimated the actual soil water content by 0.05-0.15 m 3 m -3, when using the widely applied Topp calibration function. A three-phase mixing model with a geometry parameter, a = 0.47, fit our data best. We consider mixing models to be a robust approach for calibration of TDR technique on various soils.

Abstract. The amount of nitrous oxide (N20) continues to increase in the atmosphere. Agricultural use of nitrogen fertilizers in the tropics is thought to be an important source of atmospheric N20. High frequency, highly precise measurements of the N20 flux were made with an automated system deployed in N fertilized and unfertilized agricultural plots of papaya and corn in Costa Rica for an entire corn crop growth to harvest cycle. N20 fluxes were as high as 64 ng N-N20 cm -2 h -• from fertilized versus 12 ng N-N20 cm -2 h '• from unfertilized corn and 28 ng N-N20 cm -2 h -• from fertilized versus 4.6 ng N-N20 cm -2 h '• from unfertilized papaya. Fertilized corn released more N20 than fertilized papaya over the 125 days of the crop cycle, 1.83 kg N ha -• versus 1.37 kg N ha -•. This represents a loss as N2 ¸ of 1.1 and 0.9% of the total N applied as ammonium nitrate to the corn and papaya, respectively. As has often been observed, N20 fluxes were highly variable. The fastest rates of emission were associated with fertilization and high soil moisture. A diurnal cycle in the fluxes was not evident probably due to the minimal day/night temperature fluctuations. Each chamber was measured between 509 and 523 times over the course of the experiment. This allows us to evaluate the effect on constructed mean fluxes of lowered sampling frequencies. Sampling each collar about once a day throughout the crop cycle (25% of the data set) could result in a calculated mean flux from any individual chamber that can vary by as much as 20% even though the calculated mean would probably be within 10% of the mean of the complete data set. The uncertainty increases very rapidly at lower sampling frequencies. For example, if only 10% of the data set were used which would be the equivalent of sampling every other day, a very high sampling frequency in terms of manual measurements, the calculated mean flux could vary by as much as 40% or more at any given site.