Using Slides to Test for Changes in Crown Defoliation Assessment Methods. Part I: Visual Assessment of Slides

In this study we used photographs of tree crowns to test whether the assessment methods for tree defoliation in Switzerland have changed over time. We randomly selected 24 series of slides of Norway spruce with field assessments made between 1986 and 1995. The slides were randomly arranged and assessed by three experts without prior knowledge of the year when the slide was taken or the tree number. Defoliation was assessed using the Swiss reference photo guide. Although the correlations between the field assessments and slide assessments were high (Spearman's rank correlation coefficient ranged between 0.79 and 0.83), we found significant differences between field and slide assessments (4.3 to 9% underprediction by the slide assessors) and between the slide assessments. However, no significant trends in field assessment methods could be detected. When the mean differences between field and slide assessments were subtracted, in some years, field assessors consistently underpredicted (1990, 1992) or overpredicted defoliation (1987, 1991). Defoliation tended to be overpredicted in slides taken against the light, and underpredictedfor trees with more than 25% crown overlap. We conclude that slide series can be used to detect changes in assessment methods.However, potential observer bias calls for more objectivemethods of assessment.

[1]  G. Gertner,et al.  Correlated observer errors and their effects on survey estimates of needle-leaf loss , 1995 .

[2]  J. P. Hall.  Forest health monitoring in Canada: How healthy is the boreal forest? , 1995 .

[3]  Nobuya Mizoue.  CROCO : Semi-automatic Image Analysis System for Crown Condition Assessment in Forest Health Monitoring , 2002 .

[4]  J. W. Erisman,et al.  Ten years of monitoring forest condition in Europe; studies on temporal development, spatial distribution and impacts of natural and anthropogenic stress factors , 1997 .

[5]  Geir-Harald Strand,et al.  Detection of observer bias in ongoing forest health monitoring programmes , 1996 .

[6]  Svein Solberg,et al.  Crown density assessments, control surveys and reproducibility , 1999 .

[7]  Sören Wulff.  The Accuracy of Forest Damage Assessments – Experiences from Sweden , 2002, Environmental monitoring and assessment.

[8]  J. V. Bradley.  Distribution-Free Statistical Tests , 1968 .

[9]  Nobuya Mizoue,et al.  Detecting Differences in Crown Transparency Assessments Between Countries using the Image Analysis System Croco , 2003, Environmental monitoring and assessment.

[10]  Bernhard Hanisch,et al.  Waldschäden erkennen : Fichte und Kiefer = Monitoring of forest damage : spruce and pine = Reconnaître les dommages forestiers : epicéa et pin , 1990 .

[11]  John L. Innes,et al.  Observer variation as a source of error in assessments of crown condition through time , 1995 .

[12]  Nobuya Mizoue,et al.  Using slides to test for changes in crown defoliation assessment methods part II: Application of the image analysis system CROCO , 2005, Environmental monitoring and assessment.

[13]  A. Rigling,et al.  Sanasilva-Bericht 1993. Der Zustand des Schweizer Waldes , 1994 .

[14]  L. Strand.  Crown density and fractal dimension. , 1990 .