Arts & Culture

Distribution of Landscape Types in the Global Historical Climatology Network

Distribution of Landscape Types in the Global Historical Climatology Network
of 24
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Distribution of Landscape Types inthe Global Historical ClimatologyNetwork Laure M. Montandon* Department of Geological Sciences, University of Colorado, Boulder, Colorado Souleymane Fall Department of Earth and Atmospheric Sciences, Purdue University, West Lafayette,Indiana Roger A. Pielke Sr. CIRES, University of Colorado, Boulder, Colorado Dev Niyogi Department of Earth and Atmospheric Sciences, and Department of Agronomy, andIndiana State Climate Office, Purdue University, West Lafayette, IndianaReceived 14 July 2010; accepted 13 August 2010 ABSTRACT:  The Global Historical Climate Network version 2 (GHCNv.2)surface temperature dataset is widely used for reconstructions such as theglobalaverage surface temperature (GAST) anomaly. Because land use and land cover(LULC) affect temperatures, it is important to examine the spatial distributionand the LULC representation of GHCNv.2 stations. Here, nightlight imagery,two LULC datasets, and a population and cropland historical reconstruction areused to estimate the present and historical worldwide occurrence of LULC * Corresponding author address: Laure M. Montandon, Department of Geological Sciences,University of Colorado, 2200 Colorado Ave., Boulder, CO 80309.E-mail address: Earth Interactions  d Volume 15 (2011)  d Paper No. 6  d Page 1 DOI: 10.1175/2010EI371.1 Copyright  2011, Paper 15-006; 9758 words, 8 Figures, 0 Animations, 8 Tables.  types and the number of GHCNv.2 stations within each. Results show that theGHCNv.2 station locations are biased toward urban and cropland ( . 50% sta-tions versus 18.4% of the world’s land) and past century reclaimed croplandareas (35% stations versus 3.4% land). However, widely occurring LULC suchas open shrubland, bare, snow/ice, and evergreen broadleaf forests are under-represented (14% stations versus 48.1% land), as well as nonurban areas thathave remained uncultivated in the past century (14.2% stations versus 43.2%land). Results from the temperature trends over the different landscapes confirmthat the temperature trends are different for different LULC and that theGHCNv.2 stations network might be missing on long-term larger positivetrends. This opens the possibility that the temperature increases of Earth’s landsurface in the last century would be higher than what the GHCNv.2-basedGAST analyses report. KEYWORDS:  Surfacetemperatures;Landcover;GHCN;Temperaturetrends 1. Introduction The most recent Intergovernmental Panel on Climate Change (IPCC) report(Solomon et al. 2007) identifies greenhouse gases as the major climate forcingagent responsible for the increase in temperatures. Changes in land use/land cover(LULC) and ground albedo, however, also alter the energy balance of the climate(National Research Council 2005) and are now being considered with interest (DeNoblet-Ducoudre and Pitman 2007; Pitman et al. 2009). To date, the scientific levelof understanding of the role of LULC change (and subsequent changes in albedoand surface heat capacity) on climate forcing is low relative to that of greenhousegases (Solomon et al. 2007). Many researchers, however, suggest that LULCchanges can cause large local and regional temperature changes (Pielke andAvissar 1990; Shukla et al. 1990; Kalnay and Cai 2003; Kabat et al. 2004; Limet al. 2005; Feddema et al. 2005; Pielke 2005; Chen et al. 2006; Hale et al. 2006;Hale et al. 2008; Cotton and Pielke 2007; Wichansky et al. 2008; Pielke and Niyogi2010; Mahmood et al. 2010). This effect on temperature can be the result of changes in surface roughness, vegetation amount and type, and the alteration of surface heat and moisture fluxes. LULC differ over the world and have been un-equally modified so that their distribution and the extent to which they have beenanthropogenically altered vary.Surface temperatures have been recorded at many locations around Earth formany decades. The Global Historical Climate Network (GHCN; Peterson and Vose1997) regroups these recordings in one database that is the basis for most pasttemperatures reconstructions, including the well-known global average surfacetemperature (GAST) anomaly analysis (Hansen et al. 1999; Hansen et al. 2001;Jones and Moberg 2003; Smith and Reynolds 2005). The spatial distribution of these stations is heterogeneous (Figure 1), with some areas of the world beingoverrepresented (e.g., Europe and United States) and other areas being underrep-resented (e.g., Russia, Africa, South America, and the polar regions). To correct forthis bias, in addition to other known biases such as time of observations andinstrument changes (e.g., see Pielke et al. 2000; Pielke et al. 2002; Pielke et al.2007a; Pielke et al. 2007b), the GHCN temperatures are often averaged oversmaller grid cells before being globally averaged to create a global world anomaly.The result is a more homogeneous appearing temperature record geographically. Earth Interactions  d Volume 15 (2011)  d Paper No. 6  d Page 2  However, it is unclear if the station locations adequately represent each type of landcover found on Earth as well as LULC changes during the past several decades.This is an important assessment because of the recognized impact of land-covertransformations such as urbanization, deforestation, agricultural intensification,and practices such as irrigation on climate and therefore local temperatures (Foleyet al. 2005; Roy et al. 2007; Turner et al. 2007; Fall et al. 2010a). There are alsoissues with the quality of the temperature data because of the local exposure of thesurface temperature observing sites (i.e., Davey and Pielke 2005; Pielke et al.2007a; Pielke et al. 2007b; Menne et al. 2010; Fall et al. 2010b, manuscript sub-mitted to  J. Geophys. Res. ), but we do not consider this subject in this paper.The goal of this paper is to assess the representation of the current GHCNmonthly temperature dataset [version 2 (GHCNv.2)] used to compute the GASTanomaly in terms of its representation of global LULC and their changes since the1700s. The assessment is guided from the results of many recent studies thatsuggest the LULC can impact the surface temperature trends (see Pielke et al.2007a; Mahmood et al. 2010 for a recent review). Thus, the underlying principle isthat any land-cover bias that exists in the GHCNv.2 dataset could introduce apositive (warming) or negative (cooling) bias in the GAST anomaly. This paperseeks to answer two major questions: (i) In what types of ecosystems and landcover are the GHCNv.2 stations located today? (ii) What are the historical popu-lation and cropland area changes at each GHCN v.2 station?Unlike temperatures, land-cover change is a quantity that has not been measuredcontinuously. Records of land-cover changes can rely only on historical records;inference of past human migrations; and, for the most recent years, satellite ob-servations. Historical records and past population reconstructions can be subjec-tive, and satellite observations, although useful, can have limited spatial resolutionor temporal coverage. Further, the GHCN stations themselves show significantchanges both in terms of the number of stations as well as changes in the station Figure 1. Spatial distribution of GHCNv.2 stations. Earth Interactions  d Volume 15 (2011)  d Paper No. 6  d Page 3  history (Pielke et al. 2007a). Such limitations are inherent in a global monitoringnetwork in wake of the regional development and also because the siting anddistribution are often beyond the control of a single organization. In view of theseinherent limitations, our results provide the first attempt at spatially assessing thedistribution of the GHCNv.2 dataset in term of global land-cover representation. 2. Methods 2.1. Datasets description 2.1.1. Temperature stations The GHCNv.2 temperature dataset regroups measurements from 7280 stationsfrom around the world, all with at least 10 years of data. It also includes a subset of 5206 stations that is homogeneity adjusted so that discontinuities in the tempera-ture record are corrected for biases due to changes in the instrumentation, time of observation bias, station moves, urban effects, and hand-checked exclusion of outliers from the srcinal records (Peterson and Vose 1997; Peterson 2006; Pielkeet al. 2007b). This second dataset is referred to here as the adjusted dataset. Eachstation of this dataset has at least 20 years of continuous data. The GHCNv.2temperature dataset has been used by different research institutes to compute theGASTanomaly of the past century. Theworks of the following groups are the mostcommonly cited:The Climatic Research Unit (CRU) at the University of East Anglia (Jonesand Moberg 2003);The National Climatic Data Center (NCDC) at the National Oceanic andAtmospheric Administration (Smith and Reynolds 2005); andThe Goddard Institute for Space Studies (GISS) at the National Aeronauticsand Space Administration (Hansen et al. 1999; Hansen et al. 2001).We obtained a list of stations that were used in the GAST analysis for all threedatasets (CRU, NCDC, and GISS). NCDC uses data from all stations in the srcinaladjusted GHCNv.2 dataset. GISS uses the unadjusted data from the GHCNv.2dataset, separating the U.S. stations into a distinct subdataset [U. S. HistoricalClimatology Network (USHCN)] and carrying their own adjustment independentlyon both. The CRU dataset merges some of the GHCNv.2 data with their owntemperature dataset to make it more comprehensive (Jones 1994), resulting inmany stations that overlap between their final dataset and the GHCNv.2. The mainsteps in the stations selection of each dataset are summarized in Table 1.The first step of our data preparation was to determinewhich stations in the threedatasets are GHCNv.2 stations and determine the number of spatial duplicates(stations with same geographical coordinates). For each dataset, we created twosubsets of stations. The first subset includes all stations (GHCN and non-GHCN)after removal of spatial duplicates, and it is used to compare station locations withhistorical LULC and population. The second subset includes only GHCNv.2 sta-tions. Table 2 indicates the number of stations in each subset.Typically, each temperature databasewould contain station numbers, names, andcoordinates that could be matched to GHCNv.2 values. However, in the case of the Earth Interactions  d Volume 15 (2011)  d Paper No. 6  d Page 4  Table 1. Summary of land surface temperature datasets and stations selection methodology used by three research institutescomputing GAST. ResearchinstituteLand surface temperaturedatasets usedMain steps of the analysisData exclusioncriteria No. of stationsCRU (Jones andMoberg 2003)Joint Hadley Centre/University of East Anglia Climatic ResearchUnit temperature dataset(HadCRUT) currently on itsthird version, HadCRUT3(Brohan et al. 2006), whichmerges data from manysourcesCombination of multiple recordsat each station; data spatialhomogenization over 5 83 5 8 grid;reference period: 1961–90Stations more than 5 standarddeviations from thelong-term monthly mean;duplicates with shortestrecord4138 stations (GHCNv.2and other; available online at data/landstations/)NCDC (Smith andReynolds 2005)Adjusted GHCNv.2 Data spatial homogenization over5 83 5 8 grid; reference period:1961–90Not specified All GHCNv.2 adjusted stationsGISS (Hansen et al.1999; Hansen et al.2001)Unadjusted GHCNv.2.,including data fromits U.S. subset (USHCN)through 2005; ScientificCommittee on AntarcticResearch (SCAR) datafor the AntarcticCombination of multiple records ateach station–usingthe reference station method(Peterson et al. 1998);urban adjustment: nonruralstations are adjusted so that theirlong-term trends of annual meanare as close as possible to that of the mean of the neighboringrural stations; reference period:1951–80Less than 20 yr of combinedrecord at the station; urbanstations that could not beadjusted to neighboringrural values; stations morethan 5 standard deviationsfrom the long-term monthlymean6257 GHCNv.2 stations (availableonline at gistemp/station_data/) E  ar  t    h  I    n t    er  a c t    i    o n s   d  V ol    um e 1   5    (   2   0  1  1    )     d P  a   p e r  N o . 6    d P  a   g e  5  
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks