Protected area effectiveness against land development in Spain.

Land use-land cover (LULC) changes towards artificial covers are one of the main global threats to biodiversity conservation. In this comprehensive study, we tested a number of methodological and research hypotheses, and a new covariate control technique in order to address common protected area (PA) assessment issues and accurately assess whether different PA networks have had an effect at preventing development of artificial LULCs in Spain, a highly biodiverse country that has experienced massive socioeconomic transformations in the past two decades. We used digital census data for four PA networks designated between 1990 and 2000: Nature Reserves (NRs), Nature Parks (NPs), Sites of Community Importance (SCIs) and Special Protection Areas (SPAs). We analysed the effect of explanatory variables on the ecological effectiveness of protected polygons (PPs): Legislation stringency, cummulative legal designations, management, size, age and bio-physical characteristics. A multiple Before-After-Control-Impact (BACI) semi-experimental research design was used whereby artificial land cover increase (ALCI) and proportional artificial land cover increase (PALCI) results were compared inside and outside PAs, using 1 km and 5 km buffer areas surrounding PAs as controls. LULC data were retrieved from Corine Land Cover (CLC) 1990 and 2006 data. Results from three spatial-statistical models using progressively restrictive criteria to select control areas increasingly more accurate and similar to the assessed PPs were compared. PAs were a generally effective territorial policy to prevent land development in Spain. NRs were the most effective PA category, with no new artificial land covers in the assessed period, although exact causality could not be attributed due to legal overlaps. SPAs were the least effective category, with worse ALCI data than their control areas. Legal protection was effective against land development, which was influenced by most bio-physical variables. However, cumulative legal designations and PA management did not seem to influence land development. The spatial-statistical technique used to make cases and control environmentally similar did not produce consistent outcomes and should be refined.