Assessment and monitoring of land condition in the Iberian Peninsula, 1989–2000

• Published in: Remote sensing of environment Vol. 114; no. 8; pp. 1817 - 1832
• Main Authors: del Barrio, Gabriel, Puigdefabregas, Juan, Sanjuan, Maria E., Stellmes, Marion, Ruiz, Alberto
• Format: Journal Article
• Language: English
• Published: New York, NY Elsevier Inc 16-08-2010; Elsevier
• Subjects: Animal, plant and microbial ecology; Applied geophysics; Archives; Assessment and monitoring; Assessments; Biological and medical sciences; Climate; Degradation; Desertification; Earth sciences; Earth, ocean, space; Exact sciences and technology; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; Iberian Peninsula; Internal geophysics; Land; Land condition; Land degradation; Mathematical models; MEDOKADS; Monitoring; NOAA-AVHRR; Peninsulas; Rain use efficiency; Teledetection and vegetation maps; Time–series analysis; Trends; Iberian Peninsula; Southern Europe; Europe; Europe, Iberian Peninsula; ANE, Europe, Iberian Peninsula; MED; NOAA-AVHRR; Iberian Peninsula; Assessment and monitoring; Time–series analysis; MEDOKADS; Desertification; Land condition; Land degradation; Rain use efficiency; Climatic data; Diagnosis (taxonomy); Parallel; Unit; BASIC; Implementation; Climatic condition; evaluation; Ecological damage; desertification; Procedure; Growth of error; Anthropogenic factor; biomass; rainfall; Territory; maturity; Time series; remote sensing; climate; Resilience; European Union; ecosystems; Time-series analysis; Mediterranean region; Environmental monitoring; 1989; methodology
• ISSN: 0034-4257; 1879-0704
• DOI: 10.1016/j.rse.2010.03.009

Diagnosis of land condition is a basic prerequisite for finding the degradation of a territory under climatic and human pressures leading to desertification. Ecosystemic approaches, such as the one presented here, address ecosystem maturity or resilience. They are low cost, not very prone to error propagation and well-suited to implementation on remotely sensed time–series data covering large areas. The purposes of this work were to develop a land condition surveillance methodology based on the amount of biomass produced per unit rainfall, and to test it on the Iberian Peninsula.
In this article, we propose parallel and complementary synchronic assessment and diachronic monitoring procedures to overcome the paradox of monitoring as a sequence of assessments. This is intrinsically contradictory when dealing with complex landscape mosaics, as relative estimators commonly produced for assessment are often difficult to set in a meaningful time sequence. Our approach is built on monthly time–series of two types of data, a vegetation density estimator (Green Vegetation Fraction-GVF) derived from Global Environmental Monitoring satellite archives, and corresponding interpolated climate fields. Rain Use Efficiency (RUE) is computed on two time scales to generate assessment classes. This enables detrended comparisons across different climate zones and provides automatic detection of reference areas to obtain relative RUE. The monitoring procedure uses raw GVF change rates over time and aridity in a stepwise regression to generate subclasses of discriminated trends for those drivers. The results of assessment and monitoring are then combined to yield the land condition diagnostics through explicit rules that associate their respective categories.
The approach was tested in the Iberian Peninsula for the period 1989 to 2000 using monthly GVF images derived from the 1-km MEDOKADS archive based on the NOAA-AVHRR sensors, and a corresponding archive of climate variables. The resulting land condition was validated against independent data from the Natura 2000 network of conservation reserves. In very general terms, land was found to be healthier than expected, with localised spots of ongoing degradation that were associated with current or recent intensive land use. Static or positive vegetation growth rates were detected almost everywhere, including Mediterranean areas that had undergone increased aridification during the study period. Interestingly, degrading or static trends prevailed in degraded or unusually degraded land, whereas trends to improve were most represented in land in good or unusually good condition.