HARMONDEPO harmonization of procedures of atmospheric deposition sampling Tornimbeni O., A. Marchetto, G. A. Tartari, A. Pranzo, P. Giacomotti, A. Orrù, R. Mosello C.N.R. Institute of Ecosystem Study, Verbania Pallanza, Italy. 50 m 2500 m 2 Analysis plot Control plot Large homogeneous area 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Analysis plot (50x50) Protection Stemflow collectors Throughfall collectors Snow collectors 9 samplers 0 20 40 60 80 mm H + NH 4 + Ca 2+ Mg 2+ Na + K + Alk Cl - SO 4 2- NO 3 - NT DOC % standard deviation of annual deposition 0 20 40 60 80 mm H + NH 4 + Ca 2+ Mg 2+ Na + K + Alk Cl - SO 4 2- NO 3 - NT DOC Located on a grid Randomly selected AIMS The Harmondepo project aims to evaluate the adequacy of sampling methods adopted by the Italian network CONECOFOR for level II permanent plots, in terms of sample collection and preservation. All analyses were performed on throughfall samples, taken in the permanent plots under the tree canopy. To evaluate sample stability, all throughfall samples were analysed immediately after their arrival in the laboratory, and then weekly for a month. To evaluate sampling strategies, different possible combinations and locations of the samplers in the permanent plot were simulated on the basis of individual analyses of 10 weekly samples for each of the 16 throughfall samplers. To evaluate sampling strategies, field work was performed in the EMI-1 plot (Parco Regionale Boschi di Carrega, Sala Baganza). In the above sky visions from the different throughfall sampler, it is evident that during summer the foliar canopy leads to greater differences in coverage than in winter. Three plots were sampled in the evaluation of sample stability: PIE-2 Parco Naturale Valle del Ticino, Bosco Vedro PIE-3 Parco Naturale del Devero EMI-1 Boschi Carrega SAMPLING STRATEGIES CHEMICAL STABILITY OF THE SAMPLES Winter Summer Simulating the use of 4 samplers, the relative standard deviation is greater for samples located on a grid than for random sample selection. In the case of the 9 samplers differences become very small. We conclude that 4 samplers are not enough to have a representative sampling of the area. However, 9 samplers are enough to assure that grid sampling presently used does not bias the estimated annual deposition. Throughfall samplers in a analysis plot . The use of 4 samplers introduces a relevant increase in % variance for all ionic depositions. Using 9 samplers, the increase in variance is generally markedly lower than random error. For most variables, the increase in variance due to the use of 9 samplers instead of 16 is smaller in winter than in summer, because of the greater homogeneity of the coverage. We conclude that in summer the present of 16 samplers are necessary for a correct sampling of throughfall deposition. A lower number of snow samples in winter does not imply a substantial increase in the sampling error, because of the greater homogeneity of the area. % variance % Variance CON ECO FOR Rete Nazionale Integrata Controllo Ecosistemi Forestali Poster prepared within the CONECOFOR programme, by contract with the Ministry for Agriculture and Forestry Policy – National Forest Service (Italy) in the framework of the EU Regulation (EC) No 2152/2003 (“Forest Focus”) and of the UN/ECE Program “ICP Forests”. CON ECO FOR Rete Nazionale Integrata Controllo Ecosistemi Forestali % Variance % Variance Four samples were analysed immediately after arrival in the laboratory, then stored in a dark room at 4° C for a month. All the analyses were repeated weekly. Ionic concentration randomly varied from analysis to analysis generally within a +/- 5% range around the initial values, which is comparable with the analytical error of the method used. In conclusion, throughfall samples can be stored up to one month without introducing further analytical errors, but precautions should be taken in their preservation, such as storing them in a dark cold room. VARIANCE PARTIONING FOR ANNUAL DEPOSITION was estimated by repeated random or regular selection of 4, 9 and 16 samplers from the set of 16 samplers actually analysed Winter Simulation of 9 sampler Summer Simulation of 4 samplers Simulation of 9 samplers Random error using 16 samplers Analytical error Increased variance using 9 samplers instead of the whole set of 16 Increased variance using 4 samplers instead of 9 GRID VS. RANDOM LOCATION 4 samplers LOM1 TRE1 FRI2 EMI2 TOS1 LAZ1 ABR1 CAM1 SIC1 CAL1 TOS3 TOS2 BOL1 LOM 3 LOM2 FRI1 VAL1 VEN1 MAR1 UMB1 LAZ2 SAR1 PUG1 ABR2 BAS1 LIG1 FRI2 PIE1 PIE2 PIE3 EMI1 LOM1 THF 1 THF 2 THF 3 THF 4 THF 5 THF 6 THF 7 THF 8 THF 9 THF 10 THF 11 THF 12 THF 13 THF 14 THF 15 THF 16 THF 1 THF 2 THF 3 THF 4 THF 5 THF 6 THF 7 THF 8 THF 9 THF 10 THF 11 THF 12 THF 13 THF 14 THF 15 THF 16 0% 20% 40% 60% 80% 100% mm H + NH 4 + Ca 2+ Mg 2+ Na + K + Alk Cl - SO 4 2- NO 3 + NT DOC 100% -30% -20% -10% 0% 10% 20% 30% 21/12/06 03/01/07 09/01/07 17/01/07 31/01/07 pH Cond Cl SO NO Ca Mg Na RP NT DOC Alk NH K 4 3 4 - 2- - + 2+ + 2+ + -30% -20% -10% 0% 10% 20% 30% 23/02/06 27/02/06 07/03/06 16/03/06 21/03/06 pH Cond Cl SO NO Ca Mg Na NH RP NT DOC Alk K 4 3 4 - 2- - + 2+ + 2+ + Author for correspondence: [email protected] 0% 20% 40% 60% 80% 100% mm H + NH 4 + Ca 2+ Mg 2+ Na + K + Alk Cl - SO 4 2- NO 3 - NT DOC 0% 20% 40% 60% 80% mm H + NH 4 + Ca 2+ Mg 2+ Na + K + Alk Cl - SO 4 2- NO 3 - NT DOC 0% 20% 40% 60% 80% 100% mm H + NH 4 + Ca 2+ Mg 2+ Na + K + Alk Cl - SO 4 2- NO 3 - NT DOC