Tag Archives: Tree Height

Relation of Potential Soil Carbon to Tree Height and Density across HJA

Major uncertainties exist in the geospatial distribution of terrestrial carbon (C) sources and sinks and the factors that influence soil C distribution and change. Temperate and boreal forests serve as a net sink of atmospheric CO2, so my aim is to understand how landscape features and vegetation vary across the HJ Andrews Forest (HJA) and determine how those factors influence soil C storage and CO2 release.

Research Question: How might landscape features and vegetation characteristics relate to accumulation of soil carbon, using tree height and density as a proxy for soil C?

Dataset description: I used LiDAR data from 2008 and 2011 at 0.3-0.5 m vertical resolution and at 1 m2 horizontal resolution covering the Lookout Creek Watershed and the Upper Blue River Watershed to the northern extent of HJA. These LiDAR data include a high hit model (tree tops) and a bare earth model from 2008. I also downloaded a shapefile of polygons of the reserved control areas within the HJA forest boundaries and used these areas to subset some of my later analyses.

Hypotheses:

  1. Large storm events primarily topple larger (older) trees on ridges and upper slopes. Soils in these positions are typically thinner, making them quickly saturate from precipitation and making trees in those positions more vulnerable (Lal 2005; Overby 2003). In addition, the tallest trees on ridges may be exposed to greater risk from lightning strikes, greater wind gusts and more snow accumulation, resulting in shorter trees dominating ridgelines. Shorter trees have smaller canopies and occupy less physical space than taller trees, so shorter trees occupy denser stands and as they grow taller, each individual tree occupies more space. Because of these factors, I hypothesize that shorter trees that are more closely spaced will cluster more densely along ridges than other landforms.
  2. I hypothesize that the tallest trees will be more densely clustered at low elevations along valleys because waterways carve out valleys and bring nutrients that then accumulate and build thick, carbon-rich soil. These areas tend to not suffer from the moisture limitations of ridges and steep slopes and are more protected from strong winds, so trees in these areas can maximize growth based on available solar radiation.
  3. High density young forest stands exhibit characteristics similar to mature forests, like closed canopies and high LAI. However, recently disturbed forests have higher nutrient availability than undisturbed forests which has been shown to cause a shift in C-allocation from below- to aboveground, so I expect younger stands to negatively correlate with soil C accumulation.
  4. I hypothesize that because vegetative growth is limited by solar radiation, vegetative growth on slopes that face S and W have greater exposure to solar radiation and will result in greater overall biomass (more dense stands of smaller trees, but not necessarily more tall trees). However, if the slope is too steep, I expect this pattern to diminish since trees will experience greater wind and water erosion and more tree mortality.

Note: I have yet to test hypotheses 3 and 4, but they are avenues for continuing research.

Approaches: I used a k-nearest neighbor analysis and k-means to cluster trees into 10 height classes and related each tree in the center of each height class to its distance from its 30 nearest neighboring trees of the same height class. I ran several hot spot analyses on tree heights and tree distances (tree density). Since my original hot spot analyses were subset by height class, which led to the algorithm naturally finding the arbitrary bins (upper and lower bounds) the height classes were based on, I performed new hot spot analyses on all the trees and all the tree distances within just the reserved control areas of the forest. I performed a linear regression to compare the regions of taller than expected trees to the regions of greater than expected distances between tree and Chi-squared tests independence for the hot spot analyses to compare hot spots and cold spots with variables like tree spacing, slope and elevation. Since hot spots between tree height and spacing did not overlap in all cases, I wanted to know what landscape features might explain this difference. Covariates I explored included slope, aspect, elevation, and landform. I calculated a confusion matrix between the Z-scores of height and distance for all the hot spot bins (-3,-2,-1,0,1,2,3), then further constrained the analysis to only the most extreme hot and cold spots (-3 and 3). I then compared mean height, distance, slope and elevation between the four combinations of the extreme hot and cold spots.

My objective for the final synthesis was to find regions that are higher or greater than expected for given parameters (height and tree density) and group these clusters into polygons where they overlap. I hypothesized that that soil C accumulation positively correlates with clusters of trees that are both taller and more densely spaced than expected. Conversely, I hypothesized that shorter trees that are more widely spaced negatively correlate with soil C accumulation. Therefore, overlapping hot spots of tree height and tree density and overlapping cold spots of tree height and tree density are prime areas for my planned future soil C sampling sites.

Results: I produced maps and statistical relationships between tree height and tree spacing. These included hot spot maps overlaid on elevation maps, density plots, graphical comparisons of Z-scores and many others (see previous blog posts). I found that hot spots of taller than expected trees overlapped with hot spots of greater than expected distances of trees, but not in all cases (Fig 1). When I examined a table of each of the distance hot spot bins compared with each of the height hot spot bins, I found that the most compelling correlations were between bins (0,0), (-3,-3) and (3,3) with 5%, 33% and 43% of total data, respectively. The proportion of data not covered by overlapping hot spots and by overlapping cold spots was minimal, but when mapped was visually compelling enough to warrant further exploration. I was particularly interested in areas with short trees and greater than expected distance between trees (bin -3,3) and areas with tall trees and shorter than expected distance between trees (bin 3,-3). I expect that bin (-3,3) would correlate with less soil C, so identifying those areas could be useful to sampling. I expect that bin (3,-3) would correlate with greater soil C. I can identify from Fig 2 where statistically significant hot spots and cold spots are spatially close and plan to sample heavily in those areas.

Fig 1. Tree height vs. distance between trees from hot spot analyses shows a highly linear pattern.

Fig 2. Tree density (mean distance in m between 30 closest trees) hot spot bins compared with tree height (m) hot spot bins in Lookout Mountain area. Bin comparisons (-3,-3) are areas of shorter than expected trees in more dense stands. Bin (3,-3) is taller than expected trees that are in more dense stands. Bin (3,3) is taller than expected trees in more sparsely populated stands.

 

Significance: Many of my results thus far confirm ecological phenomena that we already know to be true across forested landscapes. For example, I examined tree distance to stream and found that taller trees were more common closer to streams and less common at greater distances from streams. This makes sense with other landform features like valleys and ridges, so smaller trees tend to be along ridges (cold spots of tall trees according to my analysis) and taller trees tend to be in valleys. However, I have identified regions of taller than expected trees and regions of shorter than expected trees, and if those correlate with respectively more and less soil C, that provides evidence for LiDAR as an effective way to quantify soil C. If other landscape features that can be determined from geospatial data co-vary with tree height and/or tree density, it may be possible to quantify soil C at fine resolution. These data can be used to identify areas that have the potential to sequester more soil C and forest management could be tailored to support those regions.

Geospatial Learning: I learned a ton about ArcMap, QGIS and packages that were new to me in R like spatstat, caret, nngeo, and others. I had very little experience with and GIS before this class, so it was a steep learning curve but I’ll continue to learn as I perform this final synthesis. I learned how to perform hot spot analyses in ArcMap and export them to work with in Q. I learned how to manipulate spatial data in R and load it in Q and ArcMap for viewing. These are only a few examples of the many things I’ve learned!

Statistics Learning: I learned the limitations of hot spot analyses, with one of my criticisms being that it’s basically a smoothing function. Since the LiDAR dataset I’m using is basically a census of tree heights, running hot spot analyses is reducing the information in that dataset unnecessarily. I already knew that my data were spatially autocorrelated, so I had to take that into account for all of my analyses. I learned that the confusion matrix is great for visually discerning where a model predicts best and where it doesn’t predict well, but that the scientist must figure out the reasons for the good or poor predictions.

Because interactions among geomorphic processes, landforms, and biota occur on various temporal and spatial scales, it is necessary to understand the scale of an ecosystem process when performing spatial analyses. It is also necessary to consider the potential reasons why a particular spatial pattern did not emerge from an analysis. Reasons can range from data not being explained by expected process or multiple spatial scales of interactions, failing to use the right tool, not being at the right spatial scale, data being too coarse, failing to include the right variables or a flaw in one’s fundamental idea about how the system works. This is why it is important to (1) formulate clear, specific hypotheses before performing an analysis and (2) explore spatial patterns using multiple approaches if possible.

Sources:

Harmon M.E. (2009) Woody Detritus its Contribution to Carbon Dynamics of Old-Growth Forests: the Temporal Context. In: Wirth C., Gleixner G., Heimann M. (eds) Old-Growth Forests. Ecological Studies (Analysis and Synthesis), vol 207. Springer, Berlin, Heidelberg

Lal, R. (2005). Forest soils and carbon sequestration. Forest ecology and management, 220(1-3), 242-258.Overby et al. 2003: Impacts of natural disturbance on soil C dynamic in forest ecosystems: Soil C sources include litter fall and tree mortality, root turnover and root exudates.

Dixon, R. K., Solomon, A. M., Brown, S., Houghton, R. A., Trexier, M. C., & Wisniewski, J. (1994). Carbon pools and flux of global forest ecosystems. Science, 263(5144), 185-190.

Landscape Patterns as Predictors of Tree Height

Question: Which landscape features correspond to clusters of greater than expected trees?

Methods: I performed two Hot Spot Analyses in ArcMap; one on Hot Spots of tree height and another on hot spots of distance between trees. Both were constrained to the reserved control areas of the HJ Andrews Forest. Hot spots in tree height are regions of greater than expected tree height, while hot spots in distances between trees are regions of greater than expected distance between individual trees (more dispersed trees). Hot spots and cold spots of each analysis generally overlapped. However, hot spots between tree height and spacing did not overlap in all cases, so I wanted to know what landscape features might explain this difference. Covariates I explored included slope, aspect, elevation, and landform. Since the end goal is to find landscape features that may correlate with amount of soil carbon, I conducted this analysis with the assumption that taller trees may correlate with regions of greater soil carbon. I used the package ‘caret’ in R to calculate a  confusion matrix between the Z-scores of height and distance for all the hot spot bins (-3,-2,-1,0,1,2,3), then further constrained the analysis to only the most extreme hot and cold spots (-3 and 3). I then compared mean height, distance, slope and elevation between the four combinations of the extreme hot and cold spots (Table 1).

Results: Regions of taller than expected trees often correspond to regions of greater than expected distances between trees, which agrees with current forest growth models (Fig. 1). Hot spots of tall trees are typically in valleys and cold spots are commonly on ridges (Fig 3 & 4). When we zoom in to the Lookout Mountain area of HJ Andrews, we see that hot spots of tall trees are more concentrated in valleys and on footslopes, and cold spots are closer to mountain ridges (Fig 3). When compared with the distance hot spot map of the same area, we see that cold spots go much further down the footslopes and even into the valleys in some cases (Fig 4). So although we have evidence for a strongly linear relationship between height and distance between trees, we also have evidence that they do not fully explain each other and other landscape features are likely at play.

Fig 1. Distance Z-scores vs. Height Z-scores from hot spot analyses show a linear relationship.

Fig 2. HJ Andrews elevation with reserved control areas in orange and

inset area of Lookout Mountain hot spot maps (below)

 

Fig 3. Hot Spot Analysis showing hot spots of tree heights (tallest trees)

in the Lookout Mountain area

 

Fig 4. Hot Spot Analysis showing the greatest distance between trees

in the Lookout Mountain area

 

An elevation band that correlates with occurrences of tall trees exists up to around 1100 m, after which point number of tall trees drops off substantially (Fig. 5). Certain aspects seem to correlate with taller trees, but those relationships are harder to tease apart and I have yet to fully explore them. Greater slopes tend to correlate with shorter trees, but this relationship is not linear. There is an interesting upwards trend at slopes between 30 and 50 degrees that seems to correlate with slightly taller trees, then a big drop in mean height Z-score at slopes of 60 degrees.

Fig 5. Aspect, elevation and slope compared with Z-scores of mean height.

A comparison of Z-scores from hot spot analyses of height and distances shows that although hot spots of height and distance tightly correlate, covariates that explain them are different (mean slope and elevation). When we compare the most extreme Z-scores to one another, slope, height and distance between trees are not particularly different. Mean elevation in three categories of Z-score is similar, but mean elevation in the fourth group (>3,>3) is significantly lower. A next step is to map out these

Table 1. Comparison between the most extreme Z-scores of tree height and tree spacing.

Height Z-Score Distance Z-Score Mean Height (m) Height_SD Mean Distance (m) Distance_SD Mean Slope (m) Slope_SD Mean Elevation (m) Elevation_SD
<-3 <-3 22.8 10.5 5.1 2.4 27.9 10.5 1285 294
<-3 >3 24.5 11.2 5.6 2 26.9 4.5 1377 153
>3 <-3 34.8 7.1 5 2 31.8 5.9 1310 44
>3 >3 39.5 16.7 4.7 2.6 26.2 11 934 188

Critique: These analyses are still based on Hot Spot Analyses, so they still comes with the same criticisms as previous Hot Spot Analyses. One of these criticisms was that it’s basically a smoothing function. Since the LiDAR dataset I’m using is basically a census of tree heights, running hot spot analyses is reducing the information in that dataset unnecessarily. I have yet to map out regions that were well-predicted and poorly predicted spatially, so I cannot fully discuss the merits of the confusion matrix method.

 

Spatial Distribution of Trees by Height Class, Slope and Elevation in the HJ Andrews Forest

Guiding Questions: How do distances between trees differ depending on tree height? How does the spatial pattern of tall trees relate to the spatial pattern of slope and elevation?

Methods: I used a combination of ArcMap, QGIS and R to perform analyses and view results. I used the results of my previous distance analysis within the HJ Andrews Forest, which grouped individual trees into ten height classes and calculated the mean distance between trees within the same height class, to correlate tree spacing with other spatial phenomena. I wanted to know if hot spots in within class tree spacing correlated with hot spots in tree height, so I examined hot spots and cold spots of tree distances within each height class and compared them to tree heights, slope and elevation. Height class 1 is the shortest class of trees and height class 10 is the tallest class of trees.

I used the Hot Spot Analysis Tool in the Arc Toolbox > Spatial Statistics Tools > Mapping Clusters > Hot Spot Analysis (Getis-Ord Gi*) to perform a Hot Spot Analysis on each of the ten height classes by mean distance to the 30 closest trees of the same height class. In the context of this analysis, the interpretation of a hot spot is that it is a region of greater than expected distances between trees of the same height class. For example, in the shortest height class, 1, hot spots are regions of greater than expected mean distance between a short tree and the 30 closest short trees. Cold spots would then be regions of closer than expected mean distance between short trees.

The Hot Spot Analysis in ArcMap used a self-generated distance band of 113m for my original hot spot analysis of the global dataset (not broken up by height class), so I decided to use a distance band of 100m for each subsequent hot spot analysis. Each height class has a different number of total trees in it, so by holding the distance band constant, I hoped to avoid influence from any differences in total number of trees between height classes.

After viewing the hot spot results, I plotted the z-scores of heights for each height class against the z-scores of the distances between trees to visually examine their relationship. If both heights and distances between trees were perfectly normally distributed, one would expect a circular distribution on the density plots with a slope of zero.

I then compared the mean slopes, elevations, and standard deviation of slopes and elevation within height classes across the entire forest. Since HJA is a research forest with many different management areas, including harvested patches and research plots, I limited the next part of the analysis to only within control areas of the forest. I downloaded the most recent (2014) land use designations from the HJA data repository (http://andlter.forestry.oregonstate.edu/data/abstract.aspx?dbcode=GI008). For this analysis, I used Entity Title 3: Reserved areas (controls) within the Andrews Experimental Forest. I compared slopes and elevations within the control plots only by height class, to see if there were differences between the global dataset and the control regions of the forest.

Results:

The density plots of height z-score versus distance z-score revealed a different pattern between smaller height classes of trees and tree spacing than the relationship between larger height classes of trees and tree spacing. As we go from shorter height classes of trees to taller height classes, the density plot distributions change (Figures 1-10). There is strong evidence of positive correlation between hot spots of short trees and hot spots of distance between short trees, but from height class 6-10, there is little to no evidence of a relationship between hot spots of trees and distance between them. Tall trees are more or less distributed randomly throughout the forest.

Figure 1

 

Figure 2

 

Figure 3

 

Figure 4

 

Figure 5

 

Figure 6

 

Figure 7

 

Figure 8

 

Figure 9

 

Figure 10

There is clearly some structure to the density plots (especially in height classes 1-5), so we can assume that the trees are not randomly distributed and that there is a relationship between height and distance between trees. I compared mean and standard deviation of slope, as well as mean and standard deviation of elevation for each height class of trees (Table 1). Mean slopes do not significantly differ between height classes, so slope is likely not a main driver of tree height. However, there is some evidence that tree heights differ at lower and higher elevations, with the shortest height class of trees at a mean elevation of 1014 m, and the tallest height class at a lower mean elevation of 831 m. It’s important to note that mean elevations have large standard deviations, so the trend may not be as strong as it first appears. I wanted to know if there was more evidence for this pattern, so I calculated the same statistics for subsets of the hotspot analyses constrained to only the control areas of the forest (Table 2) to get an idea of how management may or may not influence the relationship between tree height and tree spacing throughout the forest. Mean slopes and elevations, accounting for standard deviation from the mean, do not differ significantly between the global and control datasets, meaning that the control regions are reasonable representations of the rest of the forest. To examine this further, I examined the same data for the entire forest excluding the control areas (Table 3). The same pattern holds between the three datasets; class 10 is the only class that is consistently at lower elevations across the forest. I made two density plots to display this relationship. Figure 11 shows the distribution of elevation for the shortest height class (class 1) of trees (green) versus the global dataset of all trees (red). The short trees follow the same distribution as the rest of the dataset, meaning that they are dispersed more or less evenly across elevations. Figure 12 shows the distribution of the tallest height class of trees (class 10; light blue) by elevation versus the global dataset of trees by elevation (red). This clearly shows that tall trees are not distributed at higher elevations.

Figure 11. Density of trees by elevation in height class 1 (shortest trees; green) versus global dataset of all trees (red).

 

Figure 12. Density of trees in height class 10 (tallest trees; light blue) versus global dataset of all trees (red) by elevation.

Table 1: Global dataset

Height Class Mean Slope SD Slope Mean Elevation (m) SD Elevation
1 23 11.7 1014 294
2 23 11.6 1008 291
3 25 11.6 990 295
4 25 11.5 948 291
5 25 11.3 931 293
6 26 10.9 972 291
7 27 10.5 982 250
8 26 10.6 959 221
9 25 10.7 926 207
10 23 11.2 831 197

 

Table 2: Subset of control regions

Height Class Mean Slope SD Slope Mean Elevation (m) SD Elevation
1 27 11.9 1104 317
2 27 11.5 1136 328
3 28 11 1157 329
4 28 10.8 1143 311
5 28 10.2 1133 285
6 28 10 1071 262
7 28 10 992 228
8 27 10.6 955 193
9 26 10.8 935 187
10 24 11 869 189

 

Table 3: Global dataset excluding control regions

Height Class Mean Slope SD Slope Mean Elevation (m) SD Elevation
1 22 11.4 991 284
2 22 11.4 979 273
3 24 11.6 952 272
4 24 11.5 902 266
5 24 11.5 867 265
6 25 11.3 908 290
7 26 10.7 973 267
8 25 10.5 964 242
9 23 10.4 919 221
10 22 11.2 793 198

 

Critique of the method:

A criticism of hot spot analysis is that it’s basically a smoothing function that places a focal around an area but does not account for the distribution of values within that area. So, the tallest tree in the dataset could be in cold spot (region of shorter than expected trees) and the hot spot analysis would give you no indication of that, so one may miss out on potentially useful/interesting information.

This is only a cursory look at the data and a next step is to more closely examine how slope, elevation and aspect influence distribution and height of trees, particularly within the control areas of the forest.

Point Pattern Analysis of Tree Distribution by Height in the HJ Andrews Forest

1. Given that HJ Andrew Experimental Forest is a 16,000-acre forest with manipulations spanning back to its establishment as a Long-Term Ecological Research (LTER) site in 1948, it is a highly spatially heterogeneous ecosystem. Forest harvest began in the 1950s and resulted in a mosaic of young plantation forest (~30 percent of total forest area) and old growth (~40 percent of forest) (http://andrewsforest.oregonstate.edu/). My objective is to quantify the spatial pattern of trees across the forest and eventually relate that to quantifiable landscape features.

Motivating Questions: How does the spatial pattern of trees vary across the HJ Andrews Forest? Specifically, I’m exploring the relationship between tree height and tree spacing. One specific question of interest is: How does the mean distance between trees in the same height class differ from the mean distance between a single height class of tree and all other trees? This question attempts to address the clustering vs. dispersion of trees by height.

This is an analysis of the spatial distribution of one variable, tree height, so a consideration of the internal processes that may influence the spatial distribution of this variable is necessary.

  • Microclimate caused by the clustering or dispersion of trees could be either an attraction or repulsion process. Microclimate influences relative humidity and exposure to wind, along with many other factors, so clusters of trees would tend to have different microclimate features than more dispersed trees.
  • Population and community dynamics will influence the spatial distribution of trees. The speed at which a colonizer can take over a space and competition between different colonizers influence the distribution. Aboveground and belowground tree growth adn spacing of trees will be influenced by these factors.
  • Source and sink processes in a forest may result from topographical features, like valleys and hillslopes. I expect valleys to be a source (capable of producing a surplus of organisms) sink they will tend to be in areas near streams, so not water limited and in areas that serve as catchments for nutrients, so not nutrient limited.
  • Spatial distribution of tree height certainly is different according to the scale. The spatial pattern looks different at a single tree scale compared with a 50-m scale, compared with a 5-km scale. Some of these differences are revealed by Figures 3, 4 and 5, below.

2. My approach is to use k-nearest neighbor to examine distance between a given tree and proximal trees. I created ten height classes by using kmeans to find ten cluster centers in the tree height data, then used K nearest neighbor to examine the distance between each cluster center and the 30 closest trees.

3. For this analysis, I used a LiDAR dataset of vegetation heights downloaded from the HJ Andrews online data portal (http://andlter.forestry.oregonstate.edu/data/abstract.aspx?dbcode=GI011). The selected data is from the third entry, titled, “Vegetation Height digital elevation model (DEM) from 2011 LiDAR, Upper Blue River Watershed (Oct 27th, 2011 – Nov 1st, 2011). A description of this dataset can be found here: http://andlter.forestry.oregonstate.edu/data/spatialdata/gi01103.htm.

I performed the majority of the analyses in R and used QGIS for data visualization. I used the ‘st_read’ function from the ‘sf’ package in R to read in the stem map shapefile (stem_map.shp). The stem map shapefile includes crown radii (m) and tree heights (m) as well as the point locations of trees.

Because the stem map shapefile essentially provides a census of trees within the HJA Forest, and that is (1) way too much data to deal with at one time, and (2) not useful to perform statistical analyses on since we can just look at the data mapped out and visually discern where trees are located, I decided to use kmeans to group the trees into 10 height classes, resulting in the data space being partitioned into Veronoi cells.

I used the ‘nngeo’ package to examine k-nearest neighbors relations between and within height clusters. I examined the relationship between a single tree (point) and its nearest neighbor of the same height class. I also examined the relationship between a single tree of one height class and its nearest neighbor of any height class to start to elucidate to what extent trees of similar heights cluster or are dispersed spatially.

I performed T-tests on each tree height class to test if there was a significant difference between (1) the mean distance between a tree and the next closest tree within the same height class and (2) the mean distance between a tree and the next closest tree of any height class. I calculated the mean, standard deviation and kurtosis of each height class distance (both within and between height classes).

4. Results

The table below shows the center of each of the ten height class groupings, mean distances between and within tree height classes, standard deviation and kurtosis of those means, and p-values. Results of all t-tests were significant (p<0.01), meaning that there is strong evidence that the within group mean distances are not equal to zero, so there are differences in mean distances between trees of the same height class (within group) and mean distance to the closest trees of any height class. In other words, the distance between a large tree and its nearest neighbor (of any height class) is significantly different than the distance between a large tree and its nearest neighbor within the large tree height class. The same is true of trees within and outside of the small height class, as well as each of the other ten height classes.

Table 1. Results from k-nearest neighbor analysis and t-tests

Tree Class Class Center (Height (m)) Mean Distance to 30 closest trees (m) St Dev Mean Distance Kurtosis Mean Distance Mean Distance to Closest Tree (m) St Dev Mean Mean Distance to Closest Tree Within Group (m) St Dev Group Mean P-value
1 11.8 12.5 4.4 6.4 2.1 2 4 6.3 <0.01
2 15.1 11.7 3.9 3 2.4 1.7 4.9 6.6 <0.01
3 19.9 12.6 3.4 4 3.6 1.3 6.8 6.7 <0.01
4 24.4 13.4 3.1 5.6 4 1.4 6.9 6.5 <0.01
5 29.7 14.6 2.9 3.9 4.6 1.4 8 6.9 <0.01
6 35.5 16.7 3 2.8 5.3 1.6 9.3 7.2 <0.01
7 42 18.6 2.6 4.5 6.1 1.7 10.1 7 <0.01
8 50 19.7 2.5 6.6 6.9 1.8 11.4 7.3 <0.01
9 59 20.5 2.6 3.1 7.5 1.8 13.5 8.7 <0.01
10 70 21.2 2.7 0.7 8.1 1.9 15 11 <0.01


Fig 1. Distribution of all tree heights (m) in HJ Andrews Forest.

 

 

 

Fig 2. Mean distance (m) and standard deviation (m) between trees of each of ten height classes and the next closest 30 trees within that height class. Generally, as trees get larger, mean distance between them is larger.

Fig 3. Distribution of the tallest tree class (70 m tree class) across the HJ Andrews Forest.

 

Fig 4. A representative distribution of all tree height classes on either side of a road in the HJ Andrews Forest, showing the extent of clustering and the extent of dispersion of height classes. Height classes are in ascending order from smallest class (~12m tall; Class 1) to tallest class (70m tall; Class 10).

 


Fig 5. A closer look in the same area at Fig. 4, where small trees are clustered near the road and clustered tightly together, while larger trees are more dispersed.

 

Fig 6. Mean distance between trees of the same height class and other trees of the same height class (blue) and mean distance between trees of one height class and any other tree (red). The overlapping red confidence interval with the blue points suggests that the average distance between small trees is not significantly different than distance between small trees and any trees. The general upward trend suggests that as trees get taller, the distance between them increases and variance slightly decreases.

5. Critique of the method:

The results make sense, but do not provide much more information about the actual distribution of trees (clustering vs. dispersion) than simple maps of point data, so the next step might be to examine tree heights within different management regimes. The current analysis tells me that trees are somewhat clustered by height, and that the mean distance between a tree of one height class to a tree of the same height class is, in most cases, different from the mean distance of a tree of one height class to a tree of any other height class. I’ve examined a map of different management regimes within the HJ Andrews Forest and there are clear areas of old growth, harvested areas, clearly defined plots, etc., so I would expect some of these areas to show tree clustering by height class. The patterns I found using this analysis were not as clear as I was expecting. Using kmeans and nearest neighbor analysis is a great way to start to examine the spatial relationships between and among data, but with such a large and highly varied dataset there can be shortcomings, especially when it comes to drawing any concrete conclusions.

References:

HJ Andrews Online Data Repository: http://andlter.forestry.oregonstate.edu/data/catalog/datacatalog.aspx

Johnson, S.; Lienkaemper, G. 2016. Stream network from 1997 survey and 2008 LiDAR flight, Andrews Experimental Forest. Long-Term Ecological Research. Forest Science Data Bank, Corvallis, OR. [Database]. Available: http://andlter.forestry.oregonstate.edu/data/abstract.aspx?dbcode=HF013 (10 April 2019) DOI:http://dx.doi.org/10.6073/pasta/66d98881d4eb6bb5dedcbdb60dbebafa.

Spies, T. 2016. LiDAR data (August 2008) for the Andrews Experimental Forest and Willamette National Forest study areas. Long-Term Ecological Research. Forest Science Data Bank, Corvallis, OR. [Database]. Available: http://andlter.forestry.oregonstate.edu/data/abstract.aspx?dbcode=GI010 (10 April 2019) DOI:http://dx.doi.org/10.6073/pasta/c47128d6c63dff39ee48604ecc6fabfc.

Spies, T. 2016. LiDAR data (October 2011) for the Upper Blue River Watershed, Willamette National Forest. Long-Term Ecological Research. Forest Science Data Bank, Corvallis, OR. [Database]. Available: http://andlter.forestry.oregonstate.edu/data/abstract.aspx?dbcode=GI011 (10 April 2019) DOI:http://dx.doi.org/10.6073/pasta/8e4f57bafaaad5677977dee51bb3077c.

Spies, T. 2014. Forest metrics derived from the 2008 Lidar point clouds, includes canopy closure, percentile height, and stem mapping for the Andrews Experimental Forest.. Long-Term Ecological Research. Forest Science Data Bank, Corvallis, OR. [Database]. Available: http://andlter.forestry.oregonstate.edu/data/abstract.aspx?dbcode=TV081 (10 April 2019) DOI:http://dx.doi.org/10. 6073/pasta/875e10383e8c8aee3c9a49e0155eef1d.