AZ Carbon Stores

Data

So far the data products include:

Product Downloaded (including CA) Wrangled
Xu AGBC ✅ ✅
Liu AGBC ✅ ✅
LT-GNN AGB ✅ ✅
Chopping AGB ✅ ✅
Menlove AGB ✅ ✅
ESA CCI/GlobBiomass AGB v004 ✅ ✅
GEDI L4B AGB v2.1 ✅ ✅
Rangeland Analysis Platform (RAP) AGB ✅ ✅

RAP is a model product only for rangelands (i.e. annual or perennial grasses and forbs) and is not included in comparisons right now.

Figures

Section 2.1 shows all of the data products transformed to the same extent and resolution. Section 2.2 shows the median AGB across all products. Section 2.3 shows spatial distribution of uncertainty. Section 2.4 shows the differences in AGB distribution among data products.

Observations:

  • The Liu product doesn’t include any zeros

  • The Xu product has the largest area of low but not zero values

  • The ESA product has the greatest total range of values

  • The ESA product has some very high AGB pixels surrounded by very low (maybe zero?) AGB pixels—this may be an artifact of some kind?

AGB

Arizona

Figure 1: Maps of AGB for each data product in Arizona.

California

Figure 2: Maps of AGB for each data product in California.

SRER

Figure 3: Maps of AGB for each data product in SRER

Pima County

Figure 4: Maps of AGB for each data product in Pima County

Median

Arizona

Figure 5: Median AGB across data products

California

Figure 6: Median AGB across data products

SRER

Figure 7: Median AGB across data products

Pima County

Figure 8: Median AGB across data products

Standard Deviation

Arizona

Figure 9: Map of standard deviation of each raster pixel showing areas with the highest variation among data products in yellow and areas with the least variation in dark purple. Note: the color gradient is on a log scale.

California

Figure 10: Map of standard deviation of each raster pixel showing areas with the highest variation among data products in yellow and areas with the least variation in dark purple. Note: the color gradient is on a log scale.

SRER

Figure 11: Map of standard deviation of each raster pixel showing areas with the highest variation among data products in yellow and areas with the least variation in dark purple. Note: the color gradient is on a log scale.

Pima County

Figure 12: Map of standard deviation of each raster pixel showing areas with the highest variation among data products in yellow and areas with the least variation in dark purple. Note: the color gradient is on a log scale.

Data distribution

Density ridge plots are a good compact way to visualize this.

Note

It ended up being relatively easy to just do my own kernel density estimation, so now the ridgelines are correctly truncated at min and max without having to add a point to show the range of values.

Arizona

Figure 13: Distribution of Arizona AGB values by data product.

California

Figure 14: Distribution of Arizona AGB values by data product.

SRER

Figure 15: Distribution of Arizona AGB values by data product.

Pima County

Figure 16: Distribution of Arizona AGB values by data product.

Scatter plots

Figure 17: Scatter plots showing relationship between each data product on the y-axis and the ESA product on the x-axis. Plots show a random sample of 200,000 pixels from Arizona (out of a total 35,342,827 pixels).
Figure 18: Scatter plots showing relationship between each data product on the y-axis and the ESA product on the x-axis. Plots show a random sample of 200,000 pixels from Arizona (out of a total 35,342,827 pixels).
Figure 19: Scatter plots showing relationship between each data product on the y-axis and the ESA product on the x-axis. Plots show a random sample of 200,000 pixels from Arizona (out of a total 35,342,827 pixels).
Figure 20: Scatter plots showing relationship between each data product on the y-axis and the ESA product on the x-axis. Plots show a random sample of 200,000 pixels from Arizona (out of a total 35,342,827 pixels).
Figure 21: Scatter plots showing relationship between each data product on the y-axis and the ESA product on the x-axis. Plots show a random sample of 200,000 pixels from Arizona (out of a total 35,342,827 pixels).
Figure 22: Scatter plots showing relationship between each data product on the y-axis and the ESA product on the x-axis. Plots show a random sample of 200,000 pixels from Arizona (out of a total 35,342,827 pixels).

Summary Statistics

Table 1: Summary statistics calculated on median AGB across data products for each geographical subset.
subset product min median mean max sum
AZ ESA CCI 0.00 0.00 5.84 348.00 212,900,302.00
AZ Chopping et al. 0.00 0.00 7.97 553.40 283,037,835.05
AZ Liu et al. 7.19 12.05 19.62 107.31 715,605,671.49
AZ Xu et al. 0.00 13.12 24.04 166.85 876,881,648.62
AZ LT-GNN 0.00 0.57 12.81 220.79 467,103,233.97
AZ Menlove & Healey 0.00 1.46 9.08 131.46 331,383,852.38
AZ GEDI L4B 0.00 3.86 14.38 323.79 522,118,425.15
AZ median 0.00 1.81 10.71 168.86 390,496,459.43
CA ESA CCI 0.00 0.00 42.02 559.00 2,190,200,482.00
CA Chopping et al. 0.00 0.06 45.20 990.41 2,305,074,104.07
CA Liu et al. 6.37 17.68 42.42 296.03 2,181,985,395.71
CA Xu et al. 0.00 46.05 100.16 818.49 5,213,461,856.67
CA LT-GNN 0.00 8.29 59.56 702.42 3,098,763,138.77
CA Menlove & Healey 0.00 7.23 50.42 385.14 2,628,429,901.13
CA GEDI L4B 0.00 16.09 51.77 1,330.34 2,668,091,862.80
CA median 0.00 9.99 49.76 579.85 2,594,197,630.91
SRER ESA CCI 0.00 0.00 0.30 31.00 7,419.00
SRER Chopping et al. 0.00 0.00 0.23 45.37 4,868.75
SRER Liu et al. 11.93 11.93 11.93 11.93 308,830.76
SRER Xu et al. 13.15 21.25 23.88 49.93 617,499.21
SRER LT-GNN 0.00 6.84 7.98 77.79 204,207.94
SRER Menlove & Healey 1.56 10.40 8.56 10.40 221,952.89
SRER GEDI L4B 1.24 2.71 3.77 60.91 95,225.99
SRER median 1.33 6.51 6.44 45.37 166,081.81
Pima County ESA CCI 0.00 0.00 0.94 127.00 2,675,950.00
Pima County Chopping et al. 0.00 0.00 1.42 160.50 4,060,172.95
Pima County Liu et al. 8.10 11.43 14.92 67.42 42,945,721.23
Pima County Xu et al. 0.00 10.36 13.66 89.37 39,295,863.45
Pima County LT-GNN 0.00 0.96 6.26 166.77 17,841,209.78
Pima County Menlove & Healey 0.00 0.12 1.77 22.04 5,049,265.92
Pima County GEDI L4B 0.00 2.43 6.32 186.32 18,104,745.77
Pima County median 0.00 1.26 3.72 104.00 10,681,051.77