LONG-TERM CROPPING SYSTEM EFFECTS ON MINERALIZABLE NITROGEN IN SOIL

Increasing use of N fertilizer for crop production necessitates more p recise estimates of N provided by the soil in order to prevent under- or over-fertilization and their adverse effect on plant nutrition and environmental quality. Better laboratory tests and models of N mineral ization are needed to better estimate fertilizer need. Long-term chang es in N mineralization potential may also identify changes in soil qua lity that relate to favorable sustainable agricultural practices. We c ollected samples from the 0-20 cm soil zone of treatments in five long -term (30-60 y old) experiments to determine the effect of crop rotati on, tillage, fertilizer and residue management on N mineralized during aerobic incubation, and compared results with N mineralized under fie ld conditions where possible. Soil samples were incubated in glass bot tles at 25 degrees C and -0.02 MPa for 0, 7, 14, 28 and 49 d. Soil N m ineralization from wheat (Triticum aestivum L.)-summerfallow, wheat-pe a and from wheat-wheat crop rotations were 32. 42 and 51% of that mine ralized from non-cultivated pasture soil. Nitrogen mineralized, as a f raction of the total N present, increased with increasing N applicatio n, reduction in tillage intensity and higher frequency of cropping. St ubble-mulch soils mineralized 10-20% more N than did moldboard-plowed soils. The fraction of total N mineralized increased with increasing s oil organic N content, indicating that organic N added through recent crop management practices is more labile than N in the native soil mat rix. Nitrogen mineralization in situ increased linearly as a function of past N fertilizer application, which implies that a substantial por tion of previously-applied N may be recovered slowly over time in subs equent crops and that fertilizer N needed for optimum crop yield may n ot be increasing as rapidly as expected. Nitrogen mineralized during l aboratory incubation also increased linearly with increasing N applica tion, but the rate of change differed significantly from that for in s itu N mineralization (0.0042 vs 0.0112 kg ha(-1) per kg of applied N). The difference in N mineralization rate between laboratory and in sit u experiments is not easily explained; perhaps soil processing for inc ubation altered physical access to organic N pools or caused a shift i n microbial communities in soil. The difference in N mineralization ra tes implies that laboratory incubations do not accurately reflect N mi neralization in the field, and strongly suggests that laboratory estim ates of N mineralization be interpreted with care. Previous N fertiliz ation, tillage and cropping patterns all affect N mineralization poten tial, and must be taken into consideration when estimating N fertilize r needs. (C) 1998 Elsevier Science Ltd. All rights reserved.