Estimating net primary production in 'Eucalyptus globulus' and 'Pinus pinaster' ecosystems in Portugal

The Net Primary Production (NPP) is one of the most important variables in terms of ecosystem inventory and management, because it quantifies its growth and reflects the impact of biotic and abiotic factors, which could affect it. Interest in NPP has increased recently because of the increasing interest in climate change and the need in understanding its impact on the environment. NPP is one of the most analysed key variables in this line of research. The importance of ecophysiological models, like FOREST-BGC, has also increased due to the above mentioned reasons. These types of models offer a possible methodology to test these phenomena, beyond temporal and spatial scales, not available with traditional inventory methodologies. Different methodologies to estimate NPP were tested in this project: the traditional inventory methods; the FOREST -BGC, one ecophysiological model; and the recently available NPP images (the MODIS NPP products). In terms of the traditional inventory methods, 31 and 34 sampling plots were established in a Eucalyptus globulus and Pinus pinaster stands, respectively in the North of Portugal, since those species constitute the two most important ecosystems in Portugal in terms of area. In terms of production, the results obtained from the traditional methodologies revealed that a Pinus stand is able to produce on average 14.6 ± 5.4 ton ha[sup]-1 year[sup]-1 (with 1062 trees ha[sup]-1 and 35 years old) and an Eucalyptus stand approximately 13.3 ± 4.3 ton ha[sup]-1 year[sup]-1 (with 1169 trees ha[sup]-1 and 7 years old). There are relatively few other ecosystems with the same pattern of stability and the same rates of production in Europe. In the Pinus stand litter production was extremely important and represented almost 50% while in the Eucalyptus, the arboreal component is the most important one, representing 43% of the overall NPP. Those results were extremely important and were used as reference values to compare the other methodologies used to estimate NPP. The results from those comparisons had shown that the worst results were achieved when NPP was estimated from remote sensing data exclusively. As NPP is undoubtedly a very complex variable, these results thus corroborated the claim that NPP is affected by a complex and vast number of factors and is not simply connected with the reflectance of a specific day. Out of all tested methodologies, the best results were achieved when FOREST-BGC was used to create NPP maps, after being parameterised for conditions in Portugal. Further more, the opportunity to estimate leaf area index, a key variable for the FOREST-BGC, exclusively from remotely sensed data was tested. LAI is the most important input for FOREST-BGC and this study has shown that it can be easily obtained from remotely sensed data. This broadens the range of applicability of this production model, which can now be run for smaller scale studies. Additionally several methodologies to estimate LAI were compared (measured with the ceptometer, from allometric equations) and also some corrections from the ceptometer estimations. A new correction was tested. The proposed methodology to correct the LAI ceptometer estimations constitutes one innovation in the present study, since the correction is more discriminatory than a simple constant value.