-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathss_02_location.qmd
199 lines (142 loc) · 10.9 KB
/
ss_02_location.qmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
---
output: html_document
editor_options:
chunk_output_type: console
lightbox: true
---
# Location {#sec-loc}
```{r}
library(dplyr)
library(tibble)
library(gt)
library(gtExtras)
library(sf)
library(tmap)
```
A soil description is essentially useless without a confident location. Confidence is supported by clear communication about equipment, datum and coordinate reference system, measurement units, and expected error.
## Absolute location {#sec-loc-abs}
### Recording location {#sec-loc-rec}
Record X and Y location coordinates as eastings and northings using New Zealand Transverse Mercator (EPSG:2193). For Z coordinates (elevation), use the New Zealand Vertical Datum 2016 (EPSG:1169).
X and Y precision does not need to exceed 1 m for pits, or 0.1 m for cores and auger holes. Z precision will rarely need to exceed 0.1 m.
```{r}
#| label: fig-loc
#| fig-cap: "Example location: [1586090E, 5405722N (462 m)]{.ceg}"
loc <- st_point(c(1586090, 5405722, 462), dim = "XYZ") |>
st_sfc(crs = 2193) |>
st_as_sf() |>
dplyr::mutate(label = '1586090E, 5405722N (462 m)')
nzch <- rnaturalearth::ne_countries(country = 'New Zealand', scale = 10) |>
st_transform(2193) %>%
# force match with Fig 6.1
st_crop(xmin = 952959.375186378, ymin = 4659879.67004712,
xmax = 2489543.47037023, ymax = 6237836.51954394)
tm_shape(nzch) +
tm_fill() +
tm_borders() +
tm_shape(loc) +
tm_symbols(col = 'red') +
#tm_text(text = 'label', size = 0.75, xmod = 4.25, ymod = -0.5) +
tm_compass(type='arrow') +
tm_scale_bar(breaks = c(0, 100, 200))
```
::: {.content-visible when-format='docx'}
delete me
:::
::: {#wrn-wgs84 .callout-warning collapse="true" appearance="minimal" title="Use of WGS84"}
Location sensors default to the global longitude-latitude based WGS84 datum (EPSG:4326). As a global system, its maximum accuracy is around 3 m, and coordinate accuracy also degrades over time due to continental drift. This is compensated for using a date-based 'epoch' adjustment. Because of these limitations, WGS84 should only be used when reconciling data from multiple locations on a global scale.
:::
::: {#wrn-altgeo .callout-warning collapse="true" appearance="minimal" title="Alternate geocoding systems"}
Alternate geocoding systems involving e.g. text-based encoding or nested hashes should not be used during fieldwork. These systems are often proprietary, sometimes prone to ambiguity or edge-case errors, and at their core often wrap around WGS84 anyway. They may still be useful during data analysis.
:::
### Expected Error {#sec-loc-err}
Expected horizontal error can be automatically reported by equipment, or estimated. When estimating, record in meters with a maximum precision of 0.01 (1 cm) e.g. [± 5.1 m]{.ceg}. Record expected vertical elevation error separately in the same units, as it will differ.
::: {#nte-locrev .callout-note collapse="true" appearance="minimal" title="Revising locations"}
Location data quality can sometimes be improved by revising the location data in GIS after fieldwork, with reference to e.g. satellite imagery. Where corrections are made, expected error from the measuring instrument should be overridden with a new manual estimate of expected error.
:::
### Elevation measurement {#sec-loc-abs-elev}
Parts of New Zealand still lack highly accurate elevation data. Record the elevation measurement method using the codes in @tbl-ss-elevsrc below.
```{r}
#| label: tbl-ss-elevsrc
#| tbl-cap: "Elevation data sources"
dat_ss_elevsrc <-
tribble(~Code, ~Name,
'E', 'Estimate',
'A', 'Measured by a calibrated altimeter',
'M', 'Interpolated from contour map with interval of ≤ 20 m, or extracted from similar-quality Digital Elevation Model (DEM)',
'G', 'Measured using a GNSS system with an adequate 3D fix',
'D', 'Extracted from high-accuracy DEM (Lidar or similar)')
tbl_ss_elevsrc <- gt(dat_ss_elevsrc) |>
tab_options(
column_labels.font.weight = 'bold',
heading.title.font.weight = 'bold',
table.align = 'center',
table.width = '85%'
) |>
tab_style(style = list(cell_text(color = 'red',
weight = 'bold',
align = 'center')),
locations = cells_body(columns = 1)) |>
cols_width(Code ~pct(10))
tbl_ss_elevsrc
```
The [M]{.ceg} code should be used for most non-Lidar DEMs, as those available for public use were constructed from interpolated contour data or other methods with comparable vertical accuracy [see @barringer2002; @uuemaa2020] .
Where the [A]{.ceg} code is used, the time and location of the most recent altimeter calibration should be noted.
[G]{.ceg} type elevation measurements are unlikely to be accurate if not using a [GR]{.ceg}-type sensor (see below), and may require additional conversion to the NZ Vertical Datum.
::: {#nte-dems .callout-note collapse="true" appearance="minimal" title="DEMs and location"}
Where DEMs (Digital Elevation Models) are used, they should be as contemporary as possible; New Zealand’s rate of landscape evolution is relatively high and these data sources will become less reliable with age. DEMs more than 20 years old should be accompanied by a vertical error estimate adjusted upwards for age.
:::
### Equipment {#sec-loc-eq}
Use a GNSS (Global Navigation Satellite System) with an expected horizontal accuracy of ± 3 m or better. Some applications may require centimetre accuracy. Record sensor type against the schema in @tbl-ss-locsrc.
```{r}
#| label: tbl-ss-locsrc
#| tbl-cap: "Location data source"
dat_ss_elevsrc <-
tribble(~Code, ~Name, ~Description,
'GS', 'Single-band GNSS', 'Satellite based location, single band',
'GM', 'Multi-band GNSS', 'Satellite based location, multiple band',
'GR', 'Differential GNSS', 'Satellite based location, on-ground correction',
'LM', 'Multi-sensor', 'Combined GNSS and other location data',
'NO', 'None', 'Location determined manually.')
tbl_ss_elevsrc <- gt(dat_ss_elevsrc) |>
tab_options(
column_labels.font.weight = 'bold',
heading.title.font.weight = 'bold',
table.align = 'center',
table.width = '85%'
) |>
tab_style(style = list(cell_text(color = 'red',
weight = 'bold',
align = 'center')),
locations = cells_body(columns = 1)) |>
cols_width(Code ~pct(10), Name ~ pct(20))
tbl_ss_elevsrc
```
::: {.content-visible when-format='docx'}
delete me
:::
::: {#tip-gnss .callout-tip collapse="true" appearance="minimal" title="GNSS types"}
Multi-constellation GNSS is not the same as multi-band GNSS. The former signifies access to multiple location satellite networks operated by various parties, e.g. GPS (USA), Galileo (EU), BeiDou (China). This can improve signal *consistency* due to having access to more total satellites, but not necessarily overall accuracy.
Multi-band GNSS involves receiving signals from location satellites on multiple frequency bands. The extra data from each satellite can significantly improve connection reliability and accuracy, particularly when open sky is restricted by rough terrain, dense vegetation or poor weather.
Multi-sensor GNSS involves receiving signals not just from location satellites, but from other on-ground equipment, like mobile phone towers, Wi-Fi routers and nearby personal electronic devices. These sensors can be extremely accurate in urban areas, but much less so elsewhere.
Most consumer electronic devices capable of reporting location will be of [LM]{.ceg} type by default, unless using an app that actively forces a GNSS-only connection. At the time of writing, most of these apps can still only return [GS]{.ceg}-quality data due to hardware limitations.
:::
::: {#nte-locman .callout-note collapse="true" appearance="minimal" title="Manual locations"}
Estimating location manually using e.g. topographic maps and orienteering methods remains an option, but should usually be considered a last resort.
:::
## Relative location {#sec-loc-rel}
Absolute locations are enhanced by information about what is nearby. This context is useful for checking that a site's coordinates are correct and for revisiting a site in future.
### Relative elevation {#sec-loc-relel}
Elevation above the nearest down-slope drainage (open/flowing water, swampy ground, or a closed depression) is useful for landscape interpretations. The data can also be used to group profile descriptions, for instance when identifying sets of related terraces along river systems. The drainage feature to record against must be identified by tracing a path down-slope from the observation point, and may be far from the point of observation. As such, it will normally be more efficient to estimate this parameter after fieldwork using elevation and imagery data.
Record relative elevation in meters, with a precision of no more than 0.01 (1 cm), e.g. [25.5 m]{.ceg}.
::: {#nte-rem .callout-note collapse="true" appearance="minimal" title="Relative elevation models"}
With a sufficiently detailed DEM and stream network data, a Relative Elevation Model (REM) can be constructed to estimate this parameter. The result will need to be ground-truthed.
:::
### Triangulating off local features {#sec-loc-reltri}
Absolute location data can be backed up by measuring to nearby permanent or long-term features. This is particularly useful for relocating long-term monitoring plots. For this to work well, features should be within \~50 m of the target location and sufficiently sturdy to last until at least the next expected visit. Examples include fencelines, buildings, roads, and rock outcrops.
For single-point sites, measure distance and direction using a tape and compass. Accuracy is improved by measuring to two or more features. For plot-based sites, one can measure from at least two corners to the target feature. Photographing the point/plot and reference points together is also useful.
[add a diagram here]{style="background-color:lightgreen"}
### Long-form descriptions {#sec-loc-reldesc}
For observation points that will be revisited, record practical information about how to return. This may include landholder contact details and records of previous interactions, notes about track conditions, locked gates, and potential hazards like water crossings. Spatial data recording the track from the nearest public road to the target location is particularly valuable.
### Administrative location {#sec-loc-admin}
Tagging sites with their administrative region(s) can help with discoverability (searching and filtering) in databases, and simplify information security and privacy arrangements. This does not need to be done in the field, provided accurate locations are recorded.
In New Zealand, relevant boundaries include Regional Council [@statsnz] and Territorial Authority (District Council, Stats NZ [-@statsnza]) areas. Note that these boundaries are periodically updated so need to be related to date of observation.