DISTRIBUTION AND VARIATION OF SOIL ENZYME-ACTIVITIES IN AN UNIFORM-FIELD AND THEIR INFLUENCE ON SOIL SAMPLING AND ANALYSIS

Surface soil samples of 1 ha plots (uniform soil type) of an arable la nd and of a grassland were taken according to a defined sampling patte rn. The soil samples were analysed for protease activity (from fresh s oil) and phosphatase activity (from air dried soil). The results were evaluated using different statistical techniques (classical and geosta tistical): Changes of soil enzymatic activities were observed by stora ge at + 4-degrees-C (fresh soil) and at room temperature (air dried so il). After a period of 60 days phosphatase activity decreased by 30 % in arable land and by 16 % in grassland. In the same period grassland- protease was reduced of about 24 %. An internal standard is recommende d to be carried along longer periods of analysis to correct a time dep endent decline of enzyme activities. Following sources of variance cou ld be estimated (expressed as coefficients of variation): Analysis (in cluding day to day variance): < 8 % for < 2 mm sieving and about 17 % for < 5 mm sieving. Within field: protease 14 % (arable land and grass land), phosphatase 24 % (arable land). Repeatability: < 5 % (< 2 mm si eving) and 10 % (< 5 mm sieving). Thus enzyme activities are determine d best using the sieving fraction < 2 mm of fresh soil samples. For a uniform field of 1 ha a blended sample obtained from about 30 individu al soil samples is optimal under the above conditions. By a triple ana lysis a precision of 10 % (P = 95 %) can be achieved. Enzymatic activi ty (protease) was found to be spatially dependent within the field stu died. The range of influence of one sample to another was approximatel y 40 meters, estimated by means of variogram calculation. Therefore a spatially dependent sampling can reduce the number of individual sampl es required to nine using a regular grid with 40 x 40 m.