Fig.4 Proportion of correct and erroneous trends detected surveying half of the sectors in 10 years (left) and 20 years (right) in the simulated populations of Alpine ibex in GPNP (cvy = 0.05 and cvs = 0.1). The correct magnitude correspond to the simulations in which yearly trend is estimated with less than 2% error, the biased magnitude to all the other cases in which the direction of the trend is correctly estimated but with an error higher than 2% yearly. Irrelevant trend correspond to the cases in which sample counts erroneously estimate a trend that is non-significative (P>0.05) and irrelevant trend are the simulations in which a trend in the opposite direction (increase of the population) is detected. The first bar on the left of each graph is a decrease of 50% over the time period, the rightmost a decrease of 10%