Marta Chiesi

Evaluation of statistical methods to estimate forest volume in a mediterranean region

The use of three estimation methods was investigated for mapping forest volume over a complex Mediterranean region (Tuscany, central Italy). The first two methods were based on the processing of satellite images, specifically a summer Landsat Thematic Mapper scene. From this scene, information about forest volume was extracted through a nonparametric approach [k-nearest neighbor (k-NN)] and by means of locally calibrated regressions. The last method considered, kriging, instead used only the spatial autocorrelation of tree volume relying on geostatistical principles. The experiments performed demonstrated that, at the original sampling density, the three methods produced nearly equivalent accuracies. This was no more the case when reducing the sampling density to various levels. Whereas, in fact, this reduction marginally affected the performances of the two remote-sensing-based methods, it dramatically degraded that of kriging. Additionally, the investigation showed how per-pixel estimates of error variance were obtainable also by k-NN and locally calibrated regression procedures, in analogy with the same property of kriging. Such estimated error variances were utilized to optimally integrate the outputs of the methods based on remotely sensed data and spatial autocorrelation. In all cases, the integrated estimation outperformed the single procedures. These results are relevant to develop an operational strategy for mapping forest attributes in complex Mediterranean areas

Calibration and application of FOREST-BGC in a Mediterranean area by the use of conventional and remote sensing data

The current work deals with the use in a Mediterranean environment of a simulation model of forest ecosystem processes which was originally created for temperate areas (FOREST-BGC). The model was calibrated and applied on two deciduous forest stands in Tuscany (Central Italy) by using conventional and remote sensing data as inputs. First, information on the two stands needed to initialise the model was derived from different sources, while meteorological data were extrapolated from a nearby station by an existing procedure (MT-Clim). Temporal profiles of leaf area index (LAI) were then derived both from direct ground measurement and from the processing of NOAA-AVHRR NDVI data. The model was calibrated using stand transpiration values obtained for 1997 by a sap flow method. Next, its performances were tested against the same transpiration values measured in 1998. The results obtained indicate that FOREST-BGC is capable of simulating water fluxes of Mediterranean forests when suitable LAI profiles are considered. Moreover, the derivation of these profiles from NDVI data can improve the model performance probably due to an enhanced consideration of the effects of the typical Mediterranean summer water stress. These results support the final objective of the work, which is the development of a procedure capable of integrating conventional and remote sensing data to operationally simulate water and carbon fluxes on a regional scale.

Integration of remote sensing and ecosystem modelling techniques to estimate forest net carbon uptake

Estimates of forest gross primary production (GPP) can be obtained using a parametric model (C‐Fix) that combines ground and remotely sensed data. A methodology is presented to convert these GPP estimates into values of net ecosystem exchange (NEE). The methodology is based on the use of a process model (BIOME‐BGC) that, after proper calibration, simulates all main functions of forest ecosystems at the climax condition. The estimated photosynthesis and respirations are transformed into net carbon fluxes of actual forests by using a simplified approach that relies on the difference between actual and potential stand biomass. The methodology was applied to eight forest sites in Italy where flux measurements were available and GPP estimates had been previously produced. The comparison of the obtained NEE estimates to the ground data indicates the potential of the approach and the prospects for future investigation.

Integration of high- and low-resolution satellite data to estimate pine forest productivity in a Mediterranean coastal area

The estimation of vegetation primary productivity is particularly important in fragile Mediterranean environments that are vulnerable to both natural and human-induced perturbations. The current work was aimed at using remotely sensed data taken by various sensors to infer information about a protected coastal pine forest in Tuscany (Central Italy), which could serve for driving a simplified model of carbon fluxes, C-Fix. Being based on the direct relationship between normalized difference vegetation index (NDVI) and fraction of absorbed photosynthetically active radiation (FAPAR), C-Fix uses satellite and standard meteorological data to simulate gross (GPP) and net (NPP) primary productivity of forest ecosystems. Due to the limited size of the study area, a major difficulty was in creating an NDVI dataset with suitable spatial and temporal resolutions, which was essential for the model functioning. To reach this objective, eight Thematic Mapper (TM)/Enhanced Thematic Mapper Plus (ETM+) images of two years (2000 and 2001) were merged to low-resolution NDVI estimates taken by both the Advanced Very High Resolution Radiometer (AVHRR) and VEGETATION (VGT) sensors. The C-Fix outputs for representative pine forest sites were evaluated by comparison to accurate estimates derived from a model of forest ecosystem processes previously calibrated in a similar environment (Forest-BGC). This analysis showed the potential of C-Fix for rapidly estimating GPP over wide forest areas when suitable NDVI inputs are provided. In particular, a slight superiority of VGT over AVHRR data was demonstrated, which could be reasonably attributed to the relevant higher radiometric and geometric properties. The estimation of NPP was instead quite inaccurate, due to the problematic simulation of forest respiration, which should necessarily rely on more complete modeling operations.

Simulation of Mediterranean forest carbon pools under expected environmental scenarios

Simulating the effects of possible environmental changes on the forest carbon budget requires the use of calibrated and tested models of ecosystem processes. A recently proposed simulation approach...

Monitoring water stress in Mediterranean semi-natural vegetation with satellite and meteorological data

Abstract In arid and semi-arid environments, the characterization of the inter-annual variations of the light use efficiency ɛ due to water stress still relies mostly on meteorological data. Thus the GPP estimation based on procedures exclusively driven by remote sensing data has not found yet a widespread use. In this work, the potential to characterize the water stress in semi-natural vegetation of three spectral indices (NDWI, SIWSI and NDI7) – from MODIS broad spectral bands – has been analyzed in comparison to a meteorological factor (Cws). The study comprises 70 sites (belonging to 7 different ecosystems) uniformly distributed over Tuscany, and three eddy covariance tower sites. An operational methodology, which combines meteorological and MODIS data, to characterize the inter-annual variations of ɛ due to summer water stress is proposed. Its main advantage is that it relies on existing series of meteorological data characterizing each site and allows calculating a typical Cws profile that can be “updated” ( C w s * ) for the actual conditions using MODIS spectral indices. The results confirm that the modified C w s * can be used as a proxy of water stress that does not require concurrent information on meteorological data.

Assessment of forest net primary production through the elaboration of multisource ground and remote sensing data

This paper builds on previous work by our research group which demonstrated the applicability of a parametric model, Modified C-Fix, for the monitoring of Mediterranean forests. Specifically, the model is capable of combining ground and remote sensing data to estimate forest gross primary production (GPP) on various spatial and temporal scales. Modified C-Fix is currently applied to all Italian forest areas using a previously produced data set of meteorological data and NDVI imagery descriptive of a ten-year period (1999-2008). The obtained GPP estimates are further elaborated to derive forest net primary production (NPP) averages for 20 Italian Regions. Such estimates, converted into current annual increment of standing volume (CAI) through the use of specific coefficients, are compared to the data of a recent national forest inventory (INFC). The results obtained indicate that the modelling approach tends to overestimate the ground CAI values for all forest types. The correction of a drawback in the current model implementation leads to reduce this overestimation to about 9% of the INFC increments. The possible origins of this overestimation are investigated by examining the results of previous studies and of older forest inventories. The implications of using different NPP estimation methods are finally discussed in view of assessing the forest carbon budget on a national basis.

Multi-year simulation of Mediterranean forest transpiration by the integration of NOAA-AVHRR and ancillary data

The estimation of transpiration fluxes through wide vegetated land surfaces is of great importance for the proper planning and management of environmental resources, particularly in areas where water is a main limiting factor during at least part of the growing cycle. While remotely sensed techniques cannot directly measure these fluxes, they can provide useful information on vegetation variables such as Leaf Area Index (LAI), which are functionally related to the mentioned processes. The aims of the present work were: (a) to illustrate the use of multi-temporal LAI profiles derived from National Oceanic and Atmospheric Administration Advanced Very High Resolution Radiometer (NOAA-AVHRR) Normalized Difference Vegetation Index (NDVI) data as input for a biogeochemical model (Forest-BGC) which simulates the main processes of forest vegetation (transpiration and photosynthesis); and (b) to analyse the sensitivity of the calibrated model to its main driving variables (meteorological data and NDVI-derived LAI profiles) in order to assess their relative importance for operational transpiration monitoring. In particular, the model was applied to two oak stands in the Tuscany Region (central Italy), which are representative of Mediterranean forests and for which a calibration phase had already been performed. Simulations were carried out for a 15-year period (1986–2000) using as inputs daily meteorological data and NDVI-derived monthly LAI profiles. The sensitivity of the model to both input types was then assessed through other model runs with fixed values of the two variables. The results of these experiments indicated that the remotely sensed LAI estimates are the main determinant of simulated transpirations, especially during the Mediterranean arid season (summer) when water resources are the primary limiting factor for vegetation development.

Estimating daily forest carbon fluxes using a combination of ground and remotely sensed data

Several studies have demonstrated that Monteiths approach can efficiently predict forest gross primary production (GPP), while the modeling of net ecosystem production (NEP) is more critical, requiring the additional simulation of forest respirations. The NEP of different forest ecosystems in Italy was currently simulated by the use of a remote sensing driven parametric model (modified C-Fix) and a biogeochemical model (BIOME-BGC). The outputs of the two models, which simulate forests in quasi-equilibrium conditions, are combined to estimate the carbon fluxes of actual conditions using information regarding the existing woody biomass. The estimates derived from the methodology have been tested against daily reference GPP and NEP data collected through the eddy correlation technique at five study sites in Italy. The first test concerned the theoretical validity of the simulation approach at both annual and daily time scales and was performed using optimal model drivers (i.e., collected or calibrated over the site measurements). Next, the test was repeated to assess the operational applicability of the methodology, which was driven by spatially extended data sets (i.e., data derived from existing wall-to-wall digital maps). A good estimation accuracy was generally obtained for GPP and NEP when using optimal model drivers. The use of spatially extended data sets worsens the accuracy to a varying degree, which is properly characterized. The model drivers with the most influence on the flux modeling strategy are, in increasing order of importance, forest type, soil features, meteorology, and forest woody biomass (growing stock volume).

Prediction of forest NPP in Italy by the combination of ground and remote sensing data

Our research group has recently proposed a strategy to simulate net forest carbon fluxes based on the coupling of a NDVI-driven parametric model, Modified C-Fix, and of a biogeochemical model, BIOME-BGC. The outputs of the two models are combined through the use of a proxy of ecosystem distance from equilibrium condition which accounts for the occurred disturbances. This modeling strategy is currently applied to all Italian forest areas using an available set of NDVI images and ancillary data descriptive of an 8-year period (1999–2006). The obtained estimates of forest net primary production (NPP) are first analyzed in order to assess the importance of the main model drivers on relevant spatial variability. This analysis indicates that growing stock is the most influential model driver, followed by forest type and meteorological variables. In particular, the positive influence of growing stock on NPP can be constrained by thermal and water limitations, which are most evident in the upper mountain and most southern zones, respectively. Next, the NPP estimates, aggregated over seven main forest types and twenty administrative regions in Italy, are converted into current annual increment of standing volume (CAI) by specific coefficients. The accuracy of these CAI estimates is finally assessed by comparison with the ground data collected during a recent national forest inventory. The results obtained indicate that the modeling approach tends to overestimate the ground CAI for most forest types. In particular, the overestimation is notable for forest types which are mostly managed as coppice, while it is negligible for high forests. The possible origins of these phenomena are investigated by examining the main model drivers together with the results of previous studies and of older forest inventories. The implications of using different NPP estimation methods are finally discussed in view of assessing the forest carbon budget on a national basis.