|Title||Field validation of 1930s aerial photography: What are we missing?|
|Publication Type||Journal Article|
|Year of Publication||2009|
|Authors||Browning D.M., Archer SR, Byrne A.T.|
|Journal||Journal of Arid Environments|
|ARIS Log Number||244080|
|Keywords||aboveground woody biomass, detection limites, error assessment, historid aerial photography, land cover, panchromatic, retrospective ground truth, retrospective mapping, Santa Rita, Sonoran Desert|
Aerial photography from the 1930s serves as the earliest synoptic depiction of vegetation cover. We generated a spatially explicit database of shrub (Prosopis velutina) stand structure within two 1.8 ha field plots established in 1932 to address two questions: (1) What are the detection limits of panchromatic 1936 aerial photography?, and (2) How do these influence P. velutina biomass estimates? Shrub polygons were manually digitized on 1936 imagery and linked to 1932 field measurements of P. velutina canopy area. Aboveground 1932 P. velutina biomass was estimated using a site-specific allometric relationship for field-measured canopy area. Shrub canopy detection limits on the 1936 imagery were comparable to those reported for contemporary imagery. Based on a conservative shrub size detection threshold of 3.8 m2, 5.8% of P. velutina biomass was missed. Spatial resolution (0.6 vs. 1.0 m) did not influence detection limits, but the overall accuracy of shrub cover estimates was greater on 1.0 m images. Presence of the sub-shrub Isocoma tenuisecta may also have significantly influenced estimates of P. velutina canopy area. These analyses illustrate the importance of standardizing aerial photo interpretation protocols, accounting for uncertainty estimating shrub biomass, and caution species-specific interpretations for historic aerial photography.