Nonlinear Logistic Regression of Susceptibility to Windthrow

1. Nonlinear Logistic Regression of Susceptibility to Windthrow Seminar 7 Likelihood Methods in Forest Ecology October 9th – 20th , 2006

2. Analysis of windthrow data • Traditionally: Summarize variation in degree and type of damage, across species and tree sizes, from the storm, as a whole... • A likelihood alternative: Use the spatial variation in storm intensity that occurs within a given storm to estimate parameters of functions that describe susceptibility to windthrow, as a function of variation of storm severity...

3. Categorizing Wind Damage • BINARY: Simple binary response variable (windthrown vs. not...) • CATEGORICAL: Multiple categories (uprooted, snapped,...) • ORDINAL: Ordinal categories (degree of damage): none, light, medium, heavy, complete canopy loss {usually estimated visually} • CONTINUOUS: just what the term implies, but rarely used because of the difficulties of quantifying damage in these terms...

4. Analysis of BINARY Data:Traditional Logistic Regression Consider a sample space consisting of two outcomes (A,B) where the probability that event A occurs is p • Definition: Logit = log of an odds ratio (i.e log[p/(1-p)]) • Benefits of logits • A logit is a continuous variable • Ranges from negative when p < 0.5 to positive when p > 0.5 Standard logistic regression involves fitting a linear function to the logit:

5. What if your terms are multiplicative? • Example: Assume that the probability of windthrow is a joint (multiplicative) function of • Storm severity, and • Tree size • In addition, assume that the effect of DBH is nonlinear.... • A model that incorporates these can be written as:

6. A little more detail.... • Pisjis the probability of windthrow of the jth individual of species s in plot i • DBHisj is the DBH of that individual • as, bs, and csare species-specific, estimated parameters, and • Si is the estimated storm severity in plot i • NOTE: storm severity is an arbitrary index, and was allowed to range from 0-1 But don’t you have to measure storm severity (not estimate it)?

7. Likelihood Function It couldn’t be any easier... (since the scientific model is already expressed as a probabilistic equation):.

8. An alternative, compound likelihood function and scientific model What if we assume that storm severity is not fixed within a plot (Si), but rather varies for different trees within the plot? If we are willing to assume a constant variance across all plots, this only requires estimation of a single additional parameter (the variance: s2)

9. Example: Windthrow in the Adirondacks • Highly variable damage due to: • variation within storm • topography • susceptibility of species within a stand Reference: Canham, C. D., Papaik, M. J., and Latty, E. F. 2001. Interspecific variation in susceptibility to windthrow as a function of tree size and storm severity for northern temperate tree species. Canadian Journal of Forest Research 31:1-10.

10. The dataset • Study area: 15 x 6 km area perpendicular to the storm path • 43 circular plots: 0.125 ha (19.95 m radius) censused in 1996 (20 of the 43 were in oldgrowth forests) • The plots were chosen to span a wide range of apparent damage • All trees > 10 cm DBH censused • Tallied as windthrown if uprooted or if stem was < 45o from the ground

11. Critical data requirements • Variation in storm severity across plots • Variation in DBH and species mixture within plots

12. The analysis... • 7 species comprised 97% of stems – only stems of those 7 species were included in the dataset for analysis • # parameters = 64 (43 plots + 3 parameters for each of 7 species) • Parameters estimated using simulated annealing

13. Model evaluation Numbers above bars represent the number of observations in the class The solid line is a 1:1 relationship

14. Estimating Storm Severity

15. Results: Big trees...

16. Little trees...

18. New twists • Effects of partial harvesting on risk of windthrow to residual trees • Effects of proximity to edges of clearings on risk of windthrow Research with Dave Coates in cedar-hemlock forests of interior B.C.

19. Effects of harvest intensity and proximity to edge… Equation (1): basic model – probability of windthrow is a species-specific function of tree size and storm severity: Equation (2) introduces the effect of prior harvest removal to equation (1) by adding basal area removal and assumes the effect is independent and additive Equation (3) assumes the effects of prior harvest interact with tree size: Models 1a – 3a: test models where separate c coefficients are estimated for “edge” vs. “non-edge” trees (edge = any tree within 10 m of a forest edge)

20. Other issues… • Is the risk of windthrow independent of the fate of neighboring trees? (not likely) • Should we examine spatially-explicit models that factor in the “nucleating” process of spread of windthrow gaps?…

21. Analysis for CATEGORICAL Response Variables • Extension of the binary case??: • Estimate a complete set of species-specific parameters for each of n-1 categories (assuming that the set of categories is complete and mutually exclusive...) • # of parameters required = P + (n-1)*(3*S) • Where P = # plots, S = # species, and n = # of response categories • {Is this feasible?...}

22. Analysis for ORDINAL Response Variables • The categories in this case are ranked (i.e. none, light, heavy damage) • Analysis shifts to cumulative probabilities...

23. Simple Ordinal Logistic Regression If (i.e. the probability that an observation y will be less than or equal to ordinal level Yk(k = 1.. n-1 levels), given a vector of X explanatory variables), Then simple ordinal logistic regression fits a model of the form: Remember: The probability that an event will fall into a single class k (rather than the cumulative probability) is simply

24. In our case... where and where aks, cs and bs are species specific parameters (s = 1.. m species), and Si are the estimated storm severities for the i = 1..n plots. # of parameters: M + (n-1+2)*Q, where M = # of plots, n = # of ordinal response levels, and Q = # of species

25. The Likelihood Function Stays the Same The probability that an event will fall into a single class k (rather than the cumulative probability) is simply Again, since the scientific model is already expressed as a probabilistic equation:

26. Hurricanes in Puerto Rico • Storm damage assessment in the permanent plot at the Luquillo LTER site • Hurricane Hugo - 1989 • Hurricane Georges – 1998 • Combined the data into a single analysis: 136 plots, 13 species (including 1 lumped category for “other” species), and 3 damage levels: • No or light damage • Partial damage • Complete canopy loss • Total # of parameters = 188 (15,647 trees)

27. Parameter Estimation Solving simultaneously for 188 parameters in a dataset containing > 15,000 trees takes time...

28. Model Evaluation

29. Comparison of the two storms... Statistics on variation in storm severity from Hurricanes Hugo and Georges

30. Support for the Storm Severity Parameter Estimates Support limits for the 136 estimates of storm severity were not particularly “tight” Remember that the storm severity parameter values range from 0 - 1

31. Support for the Species-specific Parameters Strength of support for the species-specific parameters was better, but still not great... Range of the 1.92 Unit Support Intervals, as a % of the parameter estimate

33. Critical assumptions • Probability of damage to a tree in Georges was independent of damage in Hugo (actually true…) • The “parallel slopes” model is reasonable • Others ?