Instrumental Validation Report

Validation of the custom LMR reconstruction against instrumental observations and the published LMRv2.1, following the methodology of PReSto2k validation.

GMST Validation Metrics (1880–2000)

Performance of the ensemble-median global mean surface temperature (GMST) against instrumental datasets over the validation window. Two metrics are reported for each reconstruction/reference pair.

Pearson correlation (R): linear association between the two time series. R measures pattern agreement but is insensitive to systematic offsets or amplitude errors.

Nash–Sutcliffe Coefficient of Efficiency (CE; Nash & Sutcliffe, 1970):
  CE = 1 − Σ(y − Ŷ)²  /  Σ(y − Ȳ)²,
where y is the observation, Ŷ is the reconstruction, and Ȳ is the mean of the observations over the validation window. CE captures pattern, amplitude, and bias simultaneously:

• CE = 1 — perfect reconstruction; residual variance is zero.
• CE = 0 — the reconstruction has the same predictive skill as simply using the observed mean; its residual variance equals the variance of the observations.
• CE < 0 — the observed mean would be a better predictor than the reconstruction.

Reference datasets: GISTEMP (NASA GISS surface temperature analysis, ERSSTv4 ocean; data.giss.nasa.gov); HadCRUT5 (Met Office Hadley Centre / CRU analysis; metoffice.gov.uk); Consensus — the arithmetic mean of the instrumental datasets listed above, taken over the years where all inputs have data. This is not an authoritative product; it is a local summary used here to smooth between-dataset differences when evaluating the reconstruction.

Spatial Validation vs GISTEMP

Grid-point R and CE between the custom reconstruction and GISTEMP over 1880–2000. Each cell's score is computed against its own 1951–1980 climatology (for anomalies) and its own time-series mean over the validation window (for the CE denominator); the geographic mean reported with each map is area-weighted.

Why the geographic-mean CE is so much lower than the GMST CE. The two numbers measure different things. At any single grid cell, most of the observed variance is local — weather noise, regional modes (ENSO, NAO, AMO, PDO) — which a paleoclimate reconstruction does not resolve. The reconstruction primarily captures the large-scale, slowly varying forced signal (volcanic, solar, anthropogenic), which is a small fraction of local variance. So per-cell CE is typically low or negative across much of the globe. When those cells are averaged into the global mean, the local noise averages towards zero and the forced signal dominates the residual variance, so the global-mean CE ends up much higher. In other words, the geographic mean of per-cell CE is not the CE of the geographic-mean time series — these are two different statistics and should not be expected to match.

Individual maps

GMST Time Series

Custom reconstruction ensemble spread compared against LMRv2.1, GISTEMP, and HadCRUT5.

GMST Ensemble Members (1000 total)

Every ensemble member plotted individually, showing the full spread of the reconstruction across all iterations and seeds.

Instrumental Period (1880-2000)

Zoomed view of ensemble members during the instrumental overlap period, with GISTEMP and HadCRUT5 overlaid.

Instrumental Period Detail

Zoomed view of the validation period (1880-2000) with ensemble spread and all reference datasets.

GMST Difference (Custom - LMRv2.1)

Year-by-year difference between the custom reconstruction and LMRv2.1 ensemble medians. Red = warmer, Blue = cooler.

Proxy Database Comparison vs PReSto2k (shared 0, custom-only 101, PReSto2k-only 683)

Comparison of the proxy records used in this reconstruction (those assimilated into the Kalman filter together with those withheld for evaluation) against the published PReSto2k reference database, matched on paleoData_TSid.

Record-selection funnel

Attrition of records from presto's initial TSID request through pipeline filtering, PSM calibration, and final assimilation.

0

Requested

240

In pickle

101

PSM-calibrated

75

Assimilated

Side-by-side statistics

Statistic	Custom (assimilated + evaluation)	PReSto2k
Number of records	101	683
Distinct archive types	1	10
Distinct proxy types	1	22
Earliest record start (Year CE)	850	850
Latest record end (Year CE)	2010	2016
Median record length (years)	235	545
Median observations per record	993	488

Compilation membership

Records counted by compilation, aggregated across all versions (hover the compilation name to see the versions present). A record is counted under every compilation its LiPD metadata claims membership of, so a record listed in both iso2k and CoralHydro2k is counted in both rows.

Source of membership data. Custom-run memberships come from each record's paleoData_inCompilationBeta field, extracted in lipd_to_pdb.py. PReSto2k records do not carry that field in presto2k_pdb.pkl; their memberships fall back to lipdverse's single paleoData_mostRecentCompilations tag (one compilation per record), so the PReSto2k counts are a lower bound. Records with no membership land in the (none) bucket.

A zero count is not a bug. Each compilation curates its own set of paleoclimate records, often with different paleoData_TSid namespaces. If presto's tsids request happens not to intersect a given compilation's records (e.g. the user requests iso2k corals, which are physically distinct samples from CoralHydro2k's corals), that compilation will show 0 even though the request listed it under compilation in query_params.json.

Compilations requested in query_params.json: Pages2kTemperature

Compilation	Custom	CSV	PReSto2k	CSV	Shared	Custom-only	PReSto2k-only
(none)	101	⬇ CSV	0	—	0	101	0
CoralHydro2k	0	—	44	⬇ CSV	0	0	44
Pages2kTemperature	0	—	636	⬇ CSV	0	0	636
iso2k	0	—	3	⬇ CSV	0	0	3

Records by archive type

Archive	Shared	PReSto2k only	Custom only	Total
Borehole	0	3	0	3
Coral	0	92	0	92
Documents	0	15	0	15
Hybrid	0	1	0	1
Ice	0	49	0	49
Lake	0	43	0	43
Marine	0	57	0	57
Other	0	0	101	101
Sclerosponge	0	5	0	5
Speleothem	0	4	0	4
Tree	0	414	0	414
Total	0	683	101	784

Detailed breakdown by proxy type

ptype	Shared	PReSto2k only	Custom only
	0	0	101
borehole	0	3	0
coral.SrCa	0	13	0
coral.calc	0	2	0
coral.d18O	0	77	0
documents	0	15	0
hybrid	0	1	0
ice.d18O	0	39	0
ice.dD	0	8	0
ice.melt	0	2	0
lake.BSi	0	2	0
lake.accumulation	0	1	0
lake.chironomid	0	8	0
lake.other	0	19	0
lake.reflectance	0	4	0
lake.varve_thickness	0	9	0
marine.MgCa	0	23	0
marine.alkenone	0	21	0
marine.other	0	13	0
sclerosponge.d18O	0	5	0
speleothem.d18O	0	4	0
tree.MXD	0	60	0
tree.TRW	0	354	0

Temporal coverage

Record counts by year, partitioned by source (custom run on top, PReSto2k below). Switch between archive-type and proxy-type (ptype) colouring.

Colour by archive type Colour by proxy type

Spatial distribution

Records exclusive to PReSto2k (683)

Records present in the PReSto2k reference but absent from the custom reconstruction (for example, archives excluded by filter_proxydb_kwargs, or TSIDs not included in the custom query).

TSID	Archive	ptype	Dataset	Years	n_obs
AB20MEN02_d18O	Coral	coral.d18O	AB20MEN02	1770–1820	592
AB20MEN03_d18O	Coral	coral.d18O	AB20MEN03	1678–1699	236
AB20MEN04_d18O	Coral	coral.d18O	AB20MEN04	1641–1684	497
AB20MEN05_d18O	Coral	coral.d18O	AB20MEN05	1532–1601	832
AB20MEN06_d18O	Coral	coral.d18O	AB20MEN06	1452–1547	1072
AB20MEN07_d18O	Coral	coral.d18O	AB20MEN07	1267–1305	466
AB20MEN08_d18O	Coral	coral.d18O	AB20MEN08	1256–1270	164
AB20MEN09_d18O	Coral	coral.d18O	AB20MEN09	1240–1262	260
Africa_001	Speleothem	speleothem.d18O	Afr-ColdAirCave.Sundqvist.2013	1635–1992	339
Africa_004	Lake	lake.other	Afr-LakeTanganyika.Tierney.2010	851–1986	60
Ant_005	Ice	ice.d18O	Ant-BerknerIsland.Mulvaney.2002	1000–1992	991
Ant_006	Ice	ice.d18O	Ant-SiteDML07.Graf.2002	1000–1994	986
Ant_007	Ice	ice.d18O	Ant-SiteDML05.Graf.2002	850–1996	1147
Ant_008	Ice	ice.d18O	Ant-SiteDML17.Graf.2002	1000–1997	984
Ant_009	Ice	ice.d18O	Ant-DomeF1993.Uemura.2014	854–1467	143
Ant_010	Ice	ice.d18O	Ant-DomeF2001.Uemura.2008	853–1875	105
Ant_011	Ice	ice.d18O	Ant-DSS.Moy.2012	850–1995	1146
Ant_012	Ice	ice.d18O	Ant-DomeC.Stenni.2001	854–1918	86
Ant_013	Ice	ice.dD	Ant-DomeC.Jouzel.2001	854–1918	86
Ant_014	Ice	ice.dD	Ant-Ferrigno.Thomas.2013	1703–2010	308
Ant_017	Ice	ice.d18O	Ant-CoastalDML.Thamban.2012	1533–1994	462
Ant_022	Ice	ice.dD	Ant-JamesRossIsland.Mulvaney.2013	850–2007	1158
Ant_023	Ice	ice.d18O	Ant-PlateauRemote.Mosley-Thompson.2013	850–1986	1137
Ant_024	Ice	ice.d18O	Ant-SipleStation.Mosley-Thompson.1990	1417–1983	567
Ant_025	Ice	ice.d18O	Ant-TALDICE.Stenni.2010	856–1990	117

⬇ Download full CSV (683 rows)

Records exclusive to the custom run (101)

Records requested in this reconstruction that do not appear in the PReSto2k reference.

TSID	Archive	ptype	Dataset	Years	n_obs
Ant-BerknerIsland.Mulvaney.2002	Other			1000–1992	993
Ant-CoastalDML.Thamban.2012	Other			1533–1994	462
Ant-DSS.Moy.2012	Other			850–1995	1146
Ant-PlateauRemote.Mosley-Thompson.2013	Other			850–1986	1137
Ant-SipleStation.Mosley-Thompson.1990	Other			1417–1983	567
Ant-SiteDML05.Graf.2002	Other			850–1996	1147
Ant-SiteDML07.Graf.2002	Other			1000–1994	995
Ant-SiteDML17.Graf.2002	Other			1000–1997	998
Ant-WDC05A.Steig.2013	Other			850–2005	1156
Ant-WDC06A.Steig.2013	Other			850–2005	23120
Arc-Agassiz.Vinther.2008	Other			850–1972	1123
Arc-Austfonna.Isaksson.2005	Other			1400–1998	554
Arc-BigRoundLake.Thomas.2009	Other			971–2003	1033
Arc-BlueLake.Bird.2008	Other			850–2000	1151
Arc-CampCentury.Fisher.1969	Other			1242–1967	726
Arc-Crete.Vinther.2010	Other			850–1973	1124
Arc-Dye.Vinther.2010	Other			850–1978	1129
Arc-GISP2.Grootes.1997	Other			850–1987	1138
Arc-GRIP.Vinther.2010	Other			850–1979	1130
Arc-Hvitarvatn.Larsen.2011	Other			850–2000	1151
Arc-LakeC2.Lamoureux.1996	Other			850–1987	1138
Arc-LakeDonardBaffinIsland.Moore.2001	Other			850–1992	1143
Arc-Lomonosovfonna.Divine.2011	Other			850–1997	1148
Arc-NGRIP1.Vinther.2006	Other			850–1995	1146
Arc-NGTB16.Schwager.1998	Other			1478–1992	515

⬇ Download full CSV (101 rows)

Records discarded during processing

Reason	Records
filtered after pickle (ptype filter or PSM calibration)	139

⬇ Download full list

← Back to results