1. Research Questions.
   • What are the spatial patterns of western juniper within the treated watershed? (The project was later expanded to address juniper density within the treated watershed, untreated watershed, and in nearby areas.)
   • What are the spatial patterns of western juniper within the study area as they relate to the spatial pattern of slope and/or aspect (as potential indicators of soil moisture) and the density of juniper in a given area (as a potential indicator of seed source)?
2. Description of dataset.
   • The study took place in the Camp Creek Paired Watershed Study (CCPWS), located in central OR. The majority of western juniper was removed from one watershed in 2005-2006 (“treated watershed), while mature western juniper is the dominant vegetation at the other watershed (“untreated watershed”).
   • Thirty-three belt transects (3m by 30m) within the treated watershed recorded the location and size of juniper within each transect.
   • Multispectral imagery were collected using a UAV for a subset of the treated watershed. Brightness values for red, green, blue, red-edge, near infrared wavelengths were used. Resolution is approximately 4 cm.
   • National Agricultural Imagery Program (NAIP) imagery from 2016 was used for the study area. Resolution of this dataset is 1 m.
   • A 30 m Digital Elevation Model (DEM) was used to determine the slope and aspect characteristics within the study area and develop a rating model as an indicator of potential soil moisture. Slope and aspect were each rated from 1-9, with areas associated with higher soil moisture (northerly aspects, flat slopes) being assigned higher values.
   • All data was projected to NAD 1983 UTM Zone 10N.
   • Supervised classification (support vector machine) was used to identify juniper within the NAIP and UAV imagery.
3. Hypotheses.
   • I expect juniper density will be dispersed unevenly throughout the study areas based on soil moisture, with high density patches of juniper being found in areas with northern aspects and/or flatter slopes. While juniper can survive in dry, rocky areas we would expect greater numbers where higher soil moisture allows for improved survival and growth. It is anticipated that patches of highest juniper density will be found in those areas with presumed highest soil moisture values based on slope and aspect characteristics, with northerly aspects and flat slopes presumed to have highest soil moisture values and areas of steep slopes with southerly aspect presumed to have the lowest soil moisture. Assuming similar soil depth and type, higher soil moisture is anticipated on northern-facing slopes compared to southerly-facing slopes because the intensity and amount of sunlight received is lower on northern slopes (in the northern hemisphere) resulting in lower soil evaporation. Higher soil moisture is presumed on flatter slopes because infiltration is assumed to be higher in these areas while higher rates of runoff are assumed on steeper slopes.
   • It is anticipated that juniper density (mature and sapling) will be higher in areas where there are higher levels of seed sources (as indicated by the juniper density surrounding each point of interest analyzed). Specifically, areas of higher density juniper saplings may be expected near higher densities of mature juniper (such as those areas near the watershed boundary). The seeds of juniper may be spread through a number of means, to include birds, winds, or small mammals.
4. Approaches.
   • Kriging and inverse distance weighted (IDW) interpolation were used with the belt transect data (represented as individual points at the start of each transect).
   • Spatial autocorrelation (Global Moran’s I) was calculated using the belt transect data (using number of juniper per transect). Spatial autocorrelation was also assessed for juniper density as it related to slope and aspect characteristics.
   • Hot spot analysis (Getis-Ord Gi) was used to assess area of juniper concentration.
   • Juniper density surrounding specific points/areas (based on NAIP imagery) was analyzed.
   • Ordinary least squares (OLS) and geographically weighted regression (GWR) were used to assess juniper density as it corresponded to slope and aspect characteristics.
   • A chi-squared test and logistic model were used to assess the rating model and classified NAIP rasters.
5. Results.
   • Based on the interpolation techniques (kriging and IDW) used areas of high and low density juniper patches were indicated within the treated watershed. However, the locations of these patches did not correspond to regions where we expected high or low soil moisture. For example, areas of highest densities of juniper density based on the transects were found in southeastern and western areas of the watershed. This may be attributed to the small dataset (n=33) used for this analysis but may also suggest that other processes (such as seed distribution via birds or the presence of other vegetation/ground cover) may influence juniper density.One hot spot was indicated using the belt transects with northwestern aspect and 8.8% slope and one cold spot was indicated with north-northwestern aspect and a 11.5% slope.
   • The hot spot analysis based on the UAV imagery indicated four hot spots, with slope values ranging from 3-11% and different aspect characteristics (northwestern, southwestern, southwestern, and southeastern).
   • The initial GWR and OLS analysis did not find that slope category and/or aspect category were good explanatory variables for juniper density based on the belt transects, based on adjusted R-squared values ranging from 0.19 to 0.28.
   • A weakly significant relationship was indicated between the rating model and the classified NAIP raster (p=0.018).
   • In order to address issues with limited variance encountered in earlier analysis, an additional GWR was conducted comparing the juniper density in a 150m buffer around 84 random points to the rating model. Outside of the buffer size, methods used in this GWR are similar to those described in previous exercises.  This improved the adjusted R-squared value for this model to 0.53.
   • 
   • The results of this analysis largely did not support the proposed hypotheses that greater juniper density would be associated with topographic characteristics associated with higher soil moisture. In situ soil moisture measurements were not used in this analysis, so differences in anticipated soil moisture associated with slope and aspect could not be compared to actual soil moisture measurements.
   • Greater densities of western juniper are evident in the untreated watershed compared to the treated watershed (as expected). However, patterns of juniper distribution within the treated watershed (as indicated by the belt transects) did not show any correlation between slope and aspect and areas of higher or lower juniper density. Additionally, those areas closest to the watershed boundary (and therefore closest to areas of dense, mature juniper) did not have greater juniper density than areas further away from the watershed boundary.
6. Significance. 
   • While results of this analysis are of limited use to land managers at this point, a similar approach with an expanded dataset may improve is utility. The use of higher resolution imagery (such as UAV imagery) may improve the identification of juniper saplings and the reliability of this analysis. The NAIP imagery used in this analysis may be sufficient for identifying mature juniper, but small (sub-meter) juniper saplings may not be accurately detected and/or misidentified. Additionally, more ground-based data (additional belt transects, etc. distributed throughout both watersheds) could provide greater information about juniper density at this study site.
   • Past research has found that juniper encroachment is associated with reduced undercanopy soil moisture [1]. The focus of this analysis was on the relationship between slope and aspect (as indicators of soil moisture) and juniper density, but further analysis may focus on how soil moisture varies with juniper density. Additionally, this analysis could also assess non-juniper vegetation density as it relates to juniper density, soil moisture, and other watershed characteristics.
   • The results suggest that other factors may be influencing juniper density in addition to slope and aspect characteristics but further analysis is needed. It should also be noted that historically western juniper was largely found in rocky outcroppings and other areas protected from fire. However, fire suppression and other factors such as livestock grazing and climate change may be contributing to an increase of juniper density [2]. Future analysis may also focus on how these factors (e.g., disturbance related to land use, etc.) are related to juniper density over larger scales.
7. My learning: Software.
   • Some difficulty encountered working when comparing raster and vector data.
   • Due to limited data set and differences in resolution, limited conclusions could be drawn from data. However, the methods could be applied to other datasets.
   • Some difficulty was encountered using ArcGIS 10.6/Pro, but issues were resolved by switching between versions.
   • Once packages and data was added, raster analysis in R-studio was relatively simple and efficient for this analysis. The R-bridge could potentially simplify this process allowing for more complicated analysis.
   • I was able to reinforce some basic skills in R and ArcGIS as well as gain exposure to new tools (such as the Ripey’s K and hot spot analysis). In particular, I was able to gain more exposure to using model builder in ArcGIS Pro for extracting and analyzing that would be time consuming and inefficient to perform otherwise.
8. My learning: Statistics.
   • I did not have previous experience with several techniques used in this analysis, to include hot spot analysis and OLS/GWR. This analysis demonstrated how these methods could be applied in the future, as well as potential limitations (e.g., small datasets, clustered data, etc.). I would be interested in applying the hot spot analysis in particular to a larger data set and to higher resolution imagery for analyzing vegetation density.
   • I have used several kriging approaches and IDW to a limited degree in the past. Similar limitations to those cited above were experienced but this is an approach that may have applicability that for in situ soil moisture measurements in the watershed.
   • In addition to those methods used in this analysis, the tutorials helped reinforce concepts related to other methods, to include: principal component analysis, Ripley’s K, and qualitative analysis using maps.
   • A major takeaway of this analysis was the potential ways that different statistical approaches can be combined to perform analysis (e.g. using a combination of OLS, Moran’s I, and GWR, etc.).

References

1. Lebron, I.; Madsen, M.D.; Chandler, D.G.; Robinson, D.A.; Wendroth, O.; Belnap, J. Ecohydrological controls on soil moisture and hydraulic conductivity within a pinyon-juniper woodland. Water Resour. Res. 2007, 43, W08422.
2. Soulé, P.T.; Knapp, P.A.; Grissino-Mayer, H.D. Human Agency, Environmental Drivers, and Western Juniper Establishment During the Late Holocene. Ecol. Appl. 2004, 14, 96–112.

1.Question asked and data used: 

How does juniper density vary with slope and aspects characteristics (as indicators of soil moisture)?

This study site for this analysis is the Camp Creek Paired Watershed Study (CCPWS), located in a semiarid region of central Oregon. This research focuses on two watersheds: one in which most juniper was removed in 2005-2006, referred to as the treated WS, and one watershed in which mature juniper is the dominant overstory species, referred to as the untreated WS. Belt transects (3m wide by 30m long) which recorded the number, height, and canopy cover of juniper within the treated WS were used for analysis. The number of juniper found in each transect was represented as a single point at the beginning of each transect. Additionally, NAIP imagery and a 30m DEM were used to assess juniper density and slope and aspect characteristics across both watersheds and in the surrounding areas.

2. Approach and tools used:

Analysis was conducted using R Studio, ArcGIS Pro, and Excel. A Chi-squared test and logistic regression were calculated in R, followed by Geographically Weighted Regression (GWR) in ArcGIS Pro. Excel was used to calculate the percentage of pixels classified as juniper within each buffer.

3. Steps used in analysis:

     A. Data preparation

              1. The following steps were conducted in a previous exercise: 1) a 30m DEM was used to extract slope and aspect characteristics, 2)slope and aspect values were each divided into nine groups, representing characteristics generally associated with increased soil moisture (e.g., flat slopes with northerly aspects were rated highest), 3)using the raster calculator the values of slope and aspect were averaged ((slope category+aspect category)/2). The raster created using this process rated each grid square from 1-9 (henceforth referred to as “rating model”). The rating model is shown below.

             2.Classification of NAIP imagery (also from a previous exercise). The NAIP imagery was classified using Support Vector Machine supervised classification. (Please note: the resolution of the NAIP imagery makes identification of juniper saplings difficult, therefore conclusions from the data should be made with caution.) The NAIP raster was resampled to the same resolution as the 30m DEM. The pixels were reclassified into a binary raster, with each pixel being either “juniper” or “not juniper”. The resampled NAIP imagery is shown below.

            3.The classified NAIP raster and rating model were projected to NAD 1983 UTM zone 10. Both rasters were masked to cover the same extent.

            4. A table was created which represented the number of juniper found per transect as a point. Using the “extract multi values to points” tool in ArcGIS the corresponding value from the rating model was assigned to each transect point.

             5. In a previous exercise, a buffer of 50m was created surrounding each of the belt transects (treated WS only). The percentage of juniper (pixels classified as juniper/total pixels) was calculated in excel and added to the belt transect table created previously.

             6. Using the “create random points” tool in ArcGIS, 96 random points were created within the study area (treated WS, untreated WS, as well as surrounding area). These points were a minimum of 150m apart.

             7. A 50m buffer was created surrounding each of the 96 random points. The percentage of juniper in each buffer was calculated as indicated above. These values were added (using the join function) to a table with the random points.

      B. Analysis

             1. The binary classified NAIP raster and rating model were loaded into R Studio and reviewed (addresses have been abbreviated):

library(raster)
library(sp)
library(rgdal)

apslop<-raster(“C:/…AspSlop_projected_clip.tif”)
juniperonly_NAIP<-raster(“C:/…NAIP_resampled_clip1.tif”)

apslop
juniperonly_NAIP
plot(apslop)
plot(juniperonly_NAIP)

s2<-stack(apslop,juniperonly_NAIP)
v2<-data.frame(na.omit(values(s2)))

             2. Chi-squared test was conducted for the rating model and classified NAIP raster (R):

chisq.test(apslop[],juniperonly_NAIP[])

             3. Logistic model, summary of results and associated plots (R):

glm.fit<-glm(juniperonly_NAIP[]~apslop[],data=v2,family=binomial)
summary(glm.fit)

plot(glm.fit)
exp(coef(glm.fit))

              4. GWR was conducted in ArcGIS for the belt transects (in treated WS), using the juniper density in the 50m buffer as the dependent variable and the rating model as the independent variable. The distance band neighborhood type and golden search neighborhood selection were used.

              5. GWR was conducted in ArcGIS using the 96 random points (in treated WS, untreated WS, and surrounding areas). The same neighborhood parameters described above were used.

              6. Global Moran’s I was calculated using the residuals for both GWR analyses. Inverse distance and the Euclidean distance method were used for analysis.

4. Overview of Results:

      A. The Chi-squared test of the binary NAIP raster (indicating juniper versus non-juniper) and the rating model raster indicated a slightly significant relationship:

Pearson’s Chi-squared test data:

apslop[] and juniperonly_NAIP[]

X-squared = 28.624, df = 15, p-value = 0.01798

      B. The logistic regression model indicated a significant relationship between the aspect and slope classification (from the rating model) and the presence or absence of juniper (dependent variable). However, the difference between the null deviance and the residual deviance suggests that this may not be an appropriate model for predicting the presence or absence of juniper. The exponent of the coefficient indicated that for every increase in aspslop rating the probability of juniper being present increased by 1.06.

Call: glm(formula = juniperonly_NAIP[] ~ apslop[], family = binomial, data = v2) Deviance Residuals:

Min 1Q Median 3Q Max

-0.6934 -0.6576 -0.6317 -0.5984 1.9537

Coefficients:

Estimate Std. Error z value Pr(>|z|)

(Intercept) -1.80732 0.11048 -16.359 < 2e-16 ***

apslop[] 0.05935 0.01948 3.046 0.00232 **

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

Null deviance: 7425.5 on 7761 degrees of freedom

Residual deviance: 7416.1 on 7760 degrees of freedom

(656 observations deleted due to missingness)

AIC: 7420.1

Number of Fisher Scoring iterations: 4

(Intercept)      apslop[]

0.164094       1.061147

      C. The GWR of treated WS dataset indicated an adjusted R-squared value of 0.28. Also, the following warning was indicated for this analysis: “At least one of your local regressions had very limited variation after applying the weights. Please use caution when interpreting the results”.  This further suggests that this model may not be indicative of the explanatory factors related to juniper density. The GWR results are provided below.

Golden Search Results

Distance Band     AICc

292.6545      222.3773

884.3284      221.4705

518.6538      216.8374

658.3291      219.2658

432.3298      217.0552

572.0050      217.5695

485.6810      216.6200

465.3026      216.6782

———————

 WARNING 110259: At least one of your local regressions had very limited variation after applying the weights. Please use caution when interpreting the results.

——————————– Analysis Details ——————————-

Number of Features:                                                             33

Dependent Variable:                      BELTTRANSECTS_50M_EXCELTOTABLE.F__JUNIPER

Explanatory Variables:  C250M_ALLSITES_EXCELTOTABLE_PROJECT.ASPSLOP_PROJECTED_CLIP

Distance Band:                                                            465.3026

———————————————————————————

–——– Model Diagnostics ———-

R2:                              0.4918

AdjR2:                           0.2845

AICc:                          216.6782

Sigma-Squared:                  28.4635

Sigma-Squared MLE:              20.4670

Effective Degrees of Freedom:   23.7290

     D. The GWR performed for the larger dataset (points distributed across both watersheds and in the surrounding areas) indicated a similar R-squared value to the GWR based in the treated WS only. A similar warning regarding the residuals was indicated for this GWR analysis. The results of this GWR are displayed below.

Golden Search Results

Distance Band     AICc

1360.4243     -98.1728

3270.5186     -84.8038

2090.0154     -90.0118

2540.9275     -87.1310

1811.3364     -92.9095

1639.1033     -95.1341

1532.6574     -96.4953

1466.8702     -97.2349

1426.2115     -97.6318

1401.0830     -97.8502

1385.5527     -97.9775

1375.9545     -98.0539

———————

WARNING 110259: At least one of your local regressions had very limited variation after applying the weights. Please use caution when interpreting the results.

—————————- Analysis Details —————————

Number of Features:                                                     96

Dependent Variable:     RANDOMPTS_50MBUFFER_JUNIPER_EXCELTOTABLE.F_JUNIPER

Explanatory Variables:            RANDOM_POINTS_105.ASPSLOP_PROJECTED_CLIP

Distance Band:                                                   1375.9545

————————————————————————-

——— Model Diagnostics ———-

R2:                              0.2932

AdjR2:                           0.1943

AICc:                          -98.0539

Sigma-Squared:                   0.0190

Sigma-Squared MLE:               0.0167

Effective Degrees of Freedom:   84.3419

     E.The Global Moran’s I for the GWR residuals from the belt transects dataset (treated WS) indicated clustering (p<0.0001) while the GWR residuals for the larger dataset (both WS with surrounding area) indicated random distribution (p=0.327).

5. Discussion/critique of methods:

The analysis used here suggests that other explanatory factors (outside of the rating model created based slope and aspect) may be in place which influence the density of juniper. The R-squared values of this analysis improved compared to previous exercises but are still relatively low. Additionally it should be noted that the presence of juniper has been correlated to reduced soil moisture in some conditions (Lebron et al., 2007), therefore slope and aspect may not be sufficient representatives of soil moisture in these watersheds.

The intent of this analysis was to develop a potential workflow that can be used with other datasets. Caution should be used in drawing any conclusions from this analysis specifically. Additionally, caution should be taken in interpreting the results of this analysis due to the characteristics of the datasets used. For instance, the resolution of the NAIP imagery makes the detection of juniper saplings difficult. Therefore the classification results may not be an accurate representation of juniper density, particularly in the treated WS. However, this process may be applied to more expansive data to include higher-resolution imagery (such as imagery collected using UAVs, etc.).

Reference:

Lebron, I., Madsen, M. D., Chandler, D. G., Robinson, D. A., Wendroth, O., & Belnap, J. (2007). Ecohydrological controls on soil moisture and hydraulic conductivity within a pinyon-juniper woodland. Water Resources Research, 43(8), W08422. https://doi.org/10.1029/2006WR005398

 

1. Question Asked: How does the density of juniper vary based on characteristics associated with soil moisture? Additionally, does this pattern vary with juniper density (as an indicator of potential seed source) surrounding each transect?

The study site for this analysis is the Camp Creek Paired Watershed Study (CCPWS) in central Oregon. The majority of juniper was removed from one watershed (“treated WS”) while one watershed is dominated by mature juniper (“untreated WS”). This analysis seeks to examine the density of juniper sapling reestablishment in the treated WS and mature juniper density in the untreated WS as it relates to indicators of soil moisture (slope, aspect, and a combination of both) and seed sources (surrounding juniper density).

2. Tools and steps used in analysis: All steps were completed in ArcGIS 10.6 or ArcGIS Pro.

Surrounding juniper density: Thirty-three belt transects were conducted in the summer of 2018 in the treated WS. The number of juniper for each transect (30m long by 3m wide) was recorded. For the purposes of this analysis, the transects were represented by a point. Supervised classification (support vector machine) was applied to National Agriculture Imagery Program (NAIP) imagery for this area and pixel-based analysis was used to estimate the density of juniper. General steps are as follows:

1. Buffers were created surrounding each transect point at 50m, 100m, and 250m using the buffer tool.
2. The zonal histogram was used to extract the pixel values within each buffer. To ensure proper counts, each buffer must be iterated separately if there is overlap between buffers.
3. A model was created in ModelBuilder to extract values from the buffers and combine values in a table. In-line variables were used to separate the data for each transect into separate tables prior to being merged.The submodel (first image) and full model (second image) are displayed below:

 

4. The percentage of juniper pixels within each buffer (number of juniper pixels/total number of pixels) was calculated using excel. This percentage was used as an estimate of surrounding juniper canopy/density within the surrounding areas.
5. Point plots were created indicating the number of juniper per transect and the corresponding juniper density.
6. The transects were divided in to two groups representing “low” counts (transects where one or fewer juniper saplings were found) and “high” counts (transects where five or greater juniper were found).

Slope and Aspect: Slope and aspect were considered to be indicators of soil moisture within the watershed. Separate approaches were used to assess if juniper density was found to vary with these characteristics: regression (using ordinary least squares and global weighted regression) and a rating model based on slope and aspect categories. General steps are as follows:

1. Using a 30m DEM, the slope and aspect values were extracted using the slope and aspect tools within ArcMap.
2. Using the calculate field function, aspect and slope were each ranked from 1-9, with areas were greater soil moisture is expected (flat and shallow slopes and northern slopes were weighted heaviest). The following parser was used (aspect values are in degrees and slope is in percent):

def Reclass(Aspect_category):

    if (Aspect_category<0 and Aspect_category>=-1):

        return 9

    if (Aspect_category>=337.5 and Aspect_category<360):

        return 8

    if (Aspect_category>=0 and Aspect_category<22.5):

        return 8

    if (Aspect_category>=22.5 and Aspect_category<67.5):

        return 7

    if (Aspect_category>=292.5 and Aspect_category<337.5):

        return 6

    if (Aspect_category>=67.5 and Aspect_category<112.5):

        return 5

    if (Aspect_category>=247.5 and Aspect_category<292.5):

        return 4

    if (Aspect_category>=112.5 and Aspect_category<157.5):

        return 3

    if (Aspect_category>=202.5 and Aspect_category<247.5):

        return 2

    if (Aspect_category>=157.5 and Aspect_category<202.5):

        return 1

def Reclass(Slope_category):

    if(Slope_category>=0 and Slope_category<2):

        return 9

    if(Slope_category>=2 and Slope_category<4):

        return 8

    if(Slope_category>=4 and Slope_category<6):

        return 7

    if(Slope_category>=6 and Slope_category<8):

        return 6

    if(Slope_category>=8 and Slope_category<10):

        return 5

    if(Slope_category>=10 and Slope_category<12):

        return 4

    if(Slope_category>=12 and Slope_category<14):

        return 3

    if(Slope_category>=14 and Slope_category<16):

        return 2

    if(Slope_category>=16):

      return 1

3. Based on the categorized values for aspect and slope, a rating model was calculated using the raster calculator to determine if areas of expected greatest soil moisture corresponded to juniper density. The formula for this rating model was: (slope category +aspect category)/2.
4. As the rating model had a lower spatial resolution and extent than the classified NAIP raster, the classified raster was resampled to be the same resolution as the rating model. The NAIP raster was masked using the extent of the rating model raster. These rasters will be used for further analysis in exercise 3. For the purposes of exercise 2, two maps were created for visual analysis. For future analysis, both rasters were projected to NAD 1983 UTM Zone 10N.
5. The hot spot analysis from exercise one was overlaid on the rating model to visually assess the patterns over a small scale.
6. An ordinary least squares (OLS) analysis was performed using the number of juniper counted for each transect as the dependent variable and the slope category, aspect category, and combination of both categories as the explanatory variables.
7. The standard residual for each OLS was used to calculate Global Moran’s I.
8. Geographically Weighted Regression (GWR) was used with the same inputs described for the OLS.

3. Overview of results.

Buffer. No clear trend was found between juniper density and the number of juniper sapling found in each transect. However, for all transects the juniper density was lower with increasing buffer distance. Results are displayed in the chart below, points are labeled by distance and “high” (corresponding to transects with 5 or more juniper) and “low” (corresponding to transects with 1 or fewer juniper:

 

Comparison of Rating Model and Classified Raster. The classified NAIP (prior to resampling and reprojection) and rating model raster are displayed below. Red pixels within the classified raster are pixels classified as juniper. For the rating model, lighter shaded areas correspond to those areas where higher soil moisture is expected based on the slope and aspect characteristics. The hot spot analysis provided limited use (as only one hot spot and one cold spot were indicated) but could be useful for additional analysis. However, the cold spot aligned with areas of expected lower soil moisture and the hot spot corresponded to areas with expected higher soil moisture. Further analysis of these two datasets will continue with exercise 3.

Regression analysis. The R-squared values for OLS for each explanatory variable(s)(slope category only, aspect category only, and a combination of slope and aspect) ranged from -0.025 to -0.060, suggesting that these characteristics were not a good predictor of juniper density (at least for the data set used in this case). The coefficient for each explanatory variable(s) ranged from -0.087 to 0.024, also indicating that this model was not a good fit. The Koenker (BP) statistic was insignificant for all cases. The p-values (p>0.86 for all cases) for Global Moran’s I for each combination of explanatory variables indicated random spatial patterns. As expected, results were similar for the GWR analysis with R-squared values ranging from -0.011 to -0.037. The GWR residual squares ranged from 180.8 to 182.3.

 

4. Discussion/critique of methods. Based on the buffer analysis, no clear patterns were indicated for repulsion/attraction based on the transect numbers and the density analysis. However, the juniper density (both in range and overall percentage) decreased for the 250m buffer regardless of the number of juniper counted per transect. These patterns are most likely related to the fact that juniper was removed from one watershed fifteen year ago and that the adjacent watershed was left untreated. This coupled with the small sample size may not be sufficient to indicate if spatial patterns between juniper density and saplings found for each transect exist. The OLS and GWR analysis indicated that slope and aspect (at least based on the data used here) were not good predictors of juniper density. For exercise 3, regression analysis of the classified raster and the rating model will take place.

The methods used here were relatively easy to apply, and can be accomplished using tools available in ArcGIS Pro. The ModelBuilder process required time to develop and some issues, particularly with the merge function, required that the tables be manually manipulated. However, ModelBuilder still was more efficient than executing the buffer and zonal histogram for each transect separately. Also, the supervised classification can easily be accomplished within ArcGIS. However, the spatial resolution of the NAIP imagery (1m) likely led to reduced accuracy of juniper sapling identification compared to images with higher spatial resolution. This was the highest resolution imagery available for the entirety of the two watersheds but this methodology could easily be applied to other data.

Overall, the methods used here may inform future analysis to be applied to other datasets. For the purpose of these exercises, very few spatial trends are evident but that may be the result of a)the sample size, b)the effects of juniper removal, and c)characteristics of the dataset (e.g., spatial resolution of the raster). In particular, I would be interested to see how these approaches would work in areas where larger-scale removal has taken place. Additionally, the use of rasters with higher spatial resolution (such as UAV-based imagery) would likely improve the classification and increase accuracy of juniper sapling detection.