Double Sampling (Subsampling)

<none>

DOUBLE SAMPLING (SUBSAMPLING)

Double sampling, also called subsampling, is a cruising methodology in which only some of the visited trees are measured.  When we use the term 'sampling', we refer to the use of Plots (or sample points) for assessing a subset of trees on a stand.  Sample plots can be fixed area or variable radius plots, and are in contrast to 100% tally cruises in which all trees are assessed at the stand level without the use of plots.  Double sampling takes this approach a step further by measuring only some trees that are visited (or assessed), and simply counting the others.  The goal of double sampling is to reduce the effort (time, cost) of cruising by reducing the number of field measurements that must be taken.  Double Sampling is typically combined with plot-based sampling but can also be used with 100% tally cruises (see Sample Tree Ratios below).  MobileMap currently supports 4 different methods of double sampling: Point Count Measure, First Tree North, Big BAF, Sample Tree Ratios. 

A key concept in double sampling is the definition of 'sample groups'.  Sample groups are typically defined by species, or by a combination of species and products.  They may also be defined by broader tree groups, like harwoods vs softwoods.  Measurements for 'measure' trees typically include, at a minimum, the species and DBH of that tree, along with enough information to calculate tree volume using the selected volume calculations for the cruise.  Even for measure trees, it is common for height information to be collected only on a subset of trees, and interpolated, or 'dubbed', using known or calculated diameter to height ratios for the remaining trees.  Tally trees, by contrast, do not have any measurements and the only information that is recorded is what is required to determine their sampling group (e.g., species and product) as well as their quantity (count). 

Below is a brief description of the double sampling approaches supported within MobileMap:

  • Point Count Measure (PCM) - Point Count Measure is perhaps the simplest method and uses two different plot types within each stand: measure plots and count plots.  On measure plots, all trees are measured with sufficient information to calculate tree volume using the selected volume calculations. PCM can be used with fixed area and variable radius plots, but not 100% tally cruises.
  • First Tree North (FTN) - First Tree North uses a single plot type.  When using FTN, the first tree of each sample group is measured on each plot.  The 'first' tree is determined by doing a sweep starting facing due North, then rotating clockwise around the plot.  The first tree encountered in each sample group is measured.  All other trees are tallied.  FTN can be used with fixed area and variable radius plots, but not 100% tally cruises.
  • Big BAF - Big BAF uses a single plot type (like FTN) but two different BAFs.  The main BAF is used to determine which trees are 'in' on the plot.  The big BAF (often 2 times the BAF of the main BAF) is used to determine which trees on the plot should be measured.  Those trees that are 'in' via the main BAF but not 'in' via the Big BAF are tallied.  Big BAF has some advantages over First Tree North in that the selection of measure trees is not random within a plot.  Big BAF will tend to select the larger trees while this may be biased, it may better represent the higher value trees that the cruiser hopes to characterize with more certainty.  Big BAF is only relevant for variable radius plots (not fixed area plots or 100% tally cruises).  More information on Big BAF can be found here https://academic.oup.com/forestry/article/90/5/649/3749606
  • Sample Tree Ratios (STR)  - Sample Tree Ratios use predefined ratios to select which trees should be measured.  Sample Tree Ratios are established for each sample group and applied across the stand (rather than at the plot level like in FTN or Big BAF).  An example ratio is 'DF:S:5' which indicates that Douglas fir sawtimber should be sampled at a ratio of 1:5, meaning that one tree should be measured out of every 5 assessed (the other 4would be tallied).  MobileMap will always flag the first tree in a sample group as a measure tree, then will use a random component when flagging additional trees.  In the previous example, it would flag tree 1 as measure, then 1 tree between trees 2 and 6, 1 tree between trees 7 and 11, etc. STR can be used with fixed area plots, variable radius plots and 100% tally cruises.

 

Data Model

In order to support sampling, several fields must be included in the cruise data model

  • Sample_Type - This field should be a text field in both the Plot and Stand feature layers that uses a Coded Value Domain that lists all double sampling methods that are supported, and an option for 'None'.  The list of accepted values are: 
    • BB = Big BAF
    • STR = Sample Tree Ratio
    • FTN = First Tree North,
    • PCM = Point Count or Measure
    • None = no sampling
  • Big_BAF - If the Big BAF method is supported, this field should be a text field that uses a Coded Value Domain that lists all of the BAF values that are appropriate for Big BAF.  Allowed values should follow the same convention
  • PCM_Ratio - If Point Count Measure method is supported, this field should be a text field that uses a Coded Value Domain that lists all of the allowed ratios in the format of '1:2', '1:3', '1:4', etc.
  • STR_Ratios - if Sample Tree Ratios are supported, this field should be a text field that will store a comma separated list of ratios (e.g., 'DF:S:5,DF:P:10,GF:XX:2,XX:XX:3')
  • Tally_Measure - This field should be a text field in the Tree feature table that uses a Coded Value Domain with the values 'M' for measure and 'T' for tally.  MobileMap will update this field automatically when it can determine if a tree should be measured or tallied.  Automatical determiniation of T or M will be possible for PCM, FTN and STR, but not for Big BAF

Settings

In addtion to these fields, there are a number of settings in MobileMap that need to be configured depending upon which Doubl Sampling methods are supported

  • Settings > Tree Sampling> General
    • Use Tree Sampling - Check box to indicate that tree sampling may be used to determine if a tree should be measured rather than tallied.  When enabled, MobileMap will alert cruiser when recording the nth tree so cruiser knows to measure DBH, height, etc.  Complete settings below to define the sampling type field and values representing the different supported sampling types
    • Stand-Level Sample Type Field - Name of field in the Stands table that stores the type of sampling (if any) should be used for a given plot.  Default field name is 'Sample_Type'. MobileMap expects that this field will use a coded value domain with the following values for supporting sampling types: BB = Big BAF, STR = Sample Tree Ratio, FTN = First Tree North, PCM = Point Count or Measure, None = no sampling.
    • Plot-Level Sample Type Field - Name of field in the Plots table that stores the type of sampling (if any) should be used for a given plot.  For 'Point Count Measure' approach, sample type at the Stand level will be used to set plots to Count or Measure.  Default field name is 'Sample_Type'. MobileMap expects that this field will use a coded value domain with the following values for supporting sampling types: BB = Big BAF, STR = Sample Tree Ratio, FTN = First Tree North, PCMC = PCM Count, PCMM = PCM Measure, Tally = all trees tallied (same as PCMC, all trees counted and not measured) None = no sampling (same as PCMM, all trees measured).
    • Tree-Level Sample Type Field Name and Values - Name of field in the Trees table (field name should be different than plot-level sample type field) that stores whether tree is a Measure tree or a Tally tree, and the values representing 'measure' and 'tally'. Use colons to separate 3 components.  Typically the values will come from a coded value domain assigned to this field.  Default field name is 'Sample' and values are 'M' and 'T' (e.g., 'Sample_Type:M:T',without quotes).  Note, cannot use non-nullable fields for fields like DHB or Height, since tally trees do not require these measurements (Species and Tally or Count can be non-nullable).
  • Settings > Tree Sampling> Sample Group Methods - FTN, STR
    • Tree Sample Group Attribute Fields - Comma separated list of fields in the tree table that define a sample group for First Tree North (FTN) and sample tree ratio (STR) subsampling.  Default value is 'Species,Product' (without quotes).
    • Sample Tree Ratios Field - Field in the Stand and/or Plot table that stores the sample tree ratios.  Default value is 'Sample_Tree_Ratios' (without quotes).
    • Default Sample Tree Ratios - Comma separated list of sample tree ratio entries.  For each sample tree ratio, enter the grouping fields and the ratio, separated by colons.  For default, use 'XX' for text or '0' for integer fields. E.g., 'DF:S:5,DF:P:10,GF:XX:2,XX:XX:3' (without quotes) indicates that every 5th Douglas fir sawtimber tree, every 10th Doug fir Pulp tree, every second Grand fir (any product), and every third tree for all other species and products should be measured and the rest of the trees should be tallied by species or species + product (according to the Sample Tree Ratio Fields above).  Each sample tree ratio included here will be included when populating a new sample tree ratio entry using the sample tree ratio editing tool.  Once created, they can be modified or augmented within that tool
  • Settings > Tree Sampling> Big BAF
    • Tree-Level BAF Field for Big BAF method - Name of field in the Trees table that stores the type of BAF that includes a given tree.  Default field name is 'BAF_Type'. MobileMap expects that this field will use a coded value domain with the following values BAF types: B = Big BAF, S = Small BAF.