Data Package Metadata   View Summary

Collecting the Data -

Plant species cover and height data are collected twice each year, spring and fall, for all sites. Spring measurements are taken in April or May when shrubs and spring annuals have reached peak biomass. Fall measurements are taken in September or October when summer annuals have reached peak biomass but prior to killing frosts. Winter measurements were taken in February before the onset of spring growth, but these winter sampling periods ended in 2007.

Vegetation data are collected on a hand-held computer. Early years used a palm top, and currently cell phones are used with a stylus.

A 1m x 1m PVC-frame is placed over the fiberglass stakes that mark the diagonal corners of each quadrat. When measuring cover it is important to stay centered over the vegetation in the quadrat to prevent errors caused by angle of view (parallax). Each PVC-frame is divided into 100 squares, marked off with nylon string. The dimensions of each square are 10cm x 10cm, and each square represents 1 percent of the total area. The cover (area) and height of each individual live (green) vegetative unit that falls within the 1m x 1m quadrat is measured. A vegetative unit consists of an individual size class (as defined by a unique cover and height) of a particular species within a quadrat. Cover is quantified by counting the number of 10cm x 10cm squares filled by each vegetative unit.

Niners and plexidecs are additional tools that help to accurately determine the cover of a vegetative unit. A niner is a small, hand-held PVC frame used to measure canopy systems or low-growing plants. The niner is also made of ½" PVC and nylon string. The interior dimensions of the niner square are 30 cm on each side, and the enclosed area is divided into nine 10 cm x 10 cm squares. Each square represents 1% of the total 1m2 quadrat cover, with the entire niner representing 30% of the total 1m2 quadrat cover. Plexidecs are used to measure vegetative units with covers < 1%. They are small, clear, ¼” plexiglass squares that are held over vegetation. A plexidec is 0.71 cm on each side, representing a cover of 0.5%. Squares are etched with dimensions of 0.1, 0.22, 0.32, and 0.5 cm. These squares delineate covers of 0.01%, 0.05%, 0.10%, and 0.25% respectively. To help relocate a plexidec dropped into vegetation, tie brightly colored surveyor's tape to a small hole drilled into one corner.

It is extremely important that cover and height measurements remain consistent over time to ensure that regressions based on this data remain valid. Field crew members should calibrate with each other to ensure that observer bias does not influence data collection.

Cover Measurements -

Cover and height measurements are made on separate vegetative units that occur within the PVC frame. A vegetative unit consists of a specific size class (as defined by a unique cover and height) of a particular species within the quadrat. This includes vegetation that is rooted outside but has foliage that hangs inside the frame.

Cover is quantified by counting the number of 10cm x 10cm squares encompassed by each vegetative unit. It is possible to obtain a total percent cover greater than 100% for a given quadrat because vegetation often overlaps, particularly shrubs (branches) and cacti (branches or pads).

The smallest cover category is 0.01%. Seedlings often have covers below 0.01%; record such seedlings as 0.01%. The next cover increments up to 1.0% are 0.05%, 0.10%, 0.25%, 0.5%, 0.75%. For covers above 1.0% up to 10.0%, the covers are rounded to the nearest 0.5%. Covers above 10.0% are rounded to the nearest 1.0%.

To increase accuracy and to reduce the size of harvested samples (see below), divide the total canopy cover of large individuals into smaller units and measure the cover and heights of each separately (exceptions may be large shrubs, cacti, and trees) . For example, an observation of black grama with a total canopy cover of 80% could be divided into several observations of smaller cover by breaking the total cover into individual clumps or groups of clumps. As a general rule, try not to record cover values that exceed 15%. During wet years, large perennial grasses can be measured up to 30%, but no higher than that.

Grasses: To determine the cover of a grass clump, envision a perimeter around the central mass or densest portion of the plant, excluding individual long leaves, wispy ends, or more open upper regions. Live tissue is frequently mixed with dead tissue in grass clumps. The goal is to measure only the plant biomass growth for the current season. In general, cover measurements should be based on green biomass. It’s not necessary to separate yellow (recently dead) biomass from green biomass on a fine scale. But large areas of yellow biomass that are most likely from a past season should not be counted in the current season’s cover. (Making this distinction requires some knowledge of the phenology of the species being measured. For example, in spring when small new green leaves of a C4 perennial grass are growing within a clump of large yellow leaves from the previous fall, count the green biomass and do not count the yellow. ELEL5 may bloom early in spring and senesce by late spring, in which case yellow biomass should be counted. Near the end of fall, biomass that was green earlier in the fall may have senesced to yellow (e.g. BOER4 stems) and should be counted.) Large areas of completely gray (long-dead) biomass within areas of live biomass should not be counted as part of the cover, but where individual gray leaves/stems are intermixed with live biomass, they can be included. Stoloniferous stems of grasses that are not rooted should be ignored. If the stem is rooted in the soil it should be recorded as a separate observation from the parent plant.

Shrubs and Sub-Shrubs: To measure dominant shrubs (e.g., Larrea tridenta, Atriplex canescens) and sub-shrubs (Gutierrezia sorothrae): Measure the cover as the perimeter of the green leaves/foliage of the plant, ignoring small open spaces (keeping in mind the 15% guideline stated above). For plants that do not have leaves, such as Ephedra torreyana, draw the perimeter around the stems/needles instead of the leaves. For Gutierrezia sarothrae, simply measure the cover and height as with any other perennial forb or grass. Do not measure dead stems or areas of dead foliage. If in doubt about whether a stem is alive, scrape the stem with your fingernail and check for the presence of green cambium. For shrubs that drop their leaves in winter, lump the branches into canopy systems and record the cover of each. It is especially important in the case of shrubs and sub-shrubs to remember to record the cover of vegetation rooted outside the quadrat but hanging inside.

Forbs: The goal in assessing forb cover is to measure all foliage produced in the current season, which usually means green foliage. Species that bloom early in spring and senesce quickly (like PLPA2, CRCR3, PHCR etc.) should be counted in spring even if they are no longer green , but shouldn’t be counted in fall unless they seem to have grown new foliage since the spring sampling. Avoid measuring gray foliage. During wet years, KAPA can be measured as a total cover if individuals cannot be identified.

Cacti: For cacti that consist of a series of pads or jointed stems (i.e., Cylindropuntia imbricata, Opuntia phaeacantha) measure the average length and width of each pad instead of cover and height. Cacti that occur as a dense ball/clump of stems (Cylindropuntia leptocaulis) are measured using the same method as shrubs, which is one total cover. Pincushion or hedgehog cacti (i.e., Echinocereus fendleri, Echinomastus intertextus, or Escobaria vivipara, also Grusonia clavata) that occur as single or clustered cylindrical stems are measured as a single cover.

Vines: Because vines often grow diffusely, measuring cover can be difficult. Each square that the vine crosses is a 0.5%. Count the squares that the vine crosses and divide by 2.

Overstory in PJ: (Quercus, Rhus, PIED, JUMO) Measure the cover as the perimeter of the green leaves/foliage of the plant above or in the quad. This number will be rough as some of the vegetation will be meters above the quad. Approximate the cover of the tree/shrub and record that value.

Height Measurements -

Height is recorded as a whole number in centimeters. Height measurements remain perpendicular to the ground even if plants are on a slope.

As of fall 2019, we’ve determined that grass and forb volume (calculated from cover and height) is not more useful than cover alone for estimating biomass of select species. Therefore, we are no longer measuring height for some grasses and forbs; a new “no-height” list will be issued and placed on the palm tops at the beginning of each field season.

Grasses: Measure the height to the top of the green foliage, ignoring inflorescence.

Forbs: If an inflorescence from the current season is present, measure the height from the base of the plant to the top of the inflorescence. If an inflorescence is not present, measure to the top of the green foliage. In the case of perennial forbs that have persistent stalks from previous years (e.g. SOEL), do not measure any old inflorescences, measure to the highest point of foliage from the current year.

Perennial Shrubs and Sub-Shrubs: Measure the height from the top of the green foliage to the ground, ignoring all bare stems.

Plants Rooted Outside but Hanging into the Quadrat: Do not measure the height to the ground. Measure only the height of the portion of the plant within the quadrat.

Overstory in PJ: Measure from the ground to the highest point of the individual tree/shrub that is either in or hanging in the quad. This is done so that the tree/shrub can be classified into strata if needed. If trees are too tall to measure to highest point, approximate by a 0.5-meter increment.

Creosote Measurements until 2013:

To measure creosote (i.e., Larrea tridenta) break the observations into two categories:

1.) Small, individual clusters of foliage on a branch (i.e., branch systems): Measure the horizontal cover of each live (i.e., green) foliage cluster, ignoring small open spaces (keeping in mind the 15% guideline stated above). Then measure the vertical "height" of each cluster from the top of the foliage to a plane created by extending a line horizontally from the bottom of the foliage. Each individual foliage cluster within a bush is considered a separate observation.

2.) Stems: Measure the length of each stem from the base to the beginning of live (i.e., green) foliage. Calculate the cumulative total of all stem measurements. This value is entered under "height" with the species as "stem" for each quadrat containing creosote. All other variable receive a default entry of "1" for creosote stem measurements.

Do not measure dead stems or areas of dead foliage. If in doubt about whether a stem is alive, scrape the stem with your fingernail and check for the presence of green cambium.

Creosote Measurements 2013 and after:

Each creosote is only measured as one total cover. Each quad that contains creosote will have one cover observation for each creosote canopy in quad.

Recording the Data -

Data are recorded using a Samsung Galaxy smartphone, commonly referred to as a palmtop. Templates specific to the season and project being sampled are downloaded onto the palmtop before the start of the season, and data are entered using the MS Excel app. A new file is created for every new site or day.

Species code, observation number, cover, height (if applicable) and count should be recorded for each observation. An observation is one specific size of a given species. If more than one individual of the same size is observed in the quad, this is recorded by increasing the count for that observation. (The same size means, the same cover and within one cm of the same height. For example, two SATR12 individuals both of cover 1, one with height 10 and the other with height 11, could be recorded as two counts of the same observation. But if that second individual has a height of 12, it needs to be a new observation.)

The observation column is essential for sorting out and correcting mislabeled plots or quadrats during the QA/QC process, so do not forget to fill in the observation column! However, it is not necessary to fill in redundant rows for site, web, plot, quad, species and/or comments; the QA/QC code will do this for you. Lower case is also fine as everything can be capitalized later. Do not use commas in any column; since the data are processed as .csv files, this will cause new columns to be made and errors to occur in the QA/QC process.

To increase accuracy and efficiency, it is recommended to:

-Make all observations for a single species before moving on to another species.

-Make all measurements on one size class within a species before moving on to another size class.

-Group observations of the same vegetative unit together and record a total count.

Rows can be inserted to add another observation of a particular species that was overlooked when the species was initially recorded.

Additional information:

Researchers involved with collecting samples/data, initials are provided in sections below for QA/QC. Chandra Tucker (CAT; 04/2014-present), Megan McClung (MAM; 04/2013-present), Stephanie Baker (SRB; 09/2010-present), John Mulhouse (JMM; 08/2009-06/2013), Amaris Swann (ALS; 08/2008-01/2013), Maya Kapoor (MLK; 08/2003 - 01/2005, 05/2010 - 03/2011), Terri Koontz (TLK; 02/2000 - 08/2003, 08/2006 - 08/2010), Yang Xia (YX; 01/2005 - 03/2010), Karen Wetherill (KRW; 02/2000 - 08/2009); Michell Thomey (MLT; 09/2005 - 08/2008), Heather Simpson (HLS; 08/2000 - 08/2002), Chris Roberts (CR; 09/2001- 08/2002), Shana Penington (SBP; 01/2000 - 08/2000), Seth Munson (SMM; 09/2002 - 06/2004), Jay McLeod (JRM; 01/2006 - 08/2006); Caleb Hickman (CRH; 09/2002 - 11/2004), Charity Hall (CLH; 01/2005 - 01/2006), Mike Friggens (MTF; 1999 - 09/2001), Tessa Edelen (MTE, 08/2004 - 08/2005).

CORE SITES -

Locating the Sampling Quadrats: Three core sites (core_blue, core_black, core_creosote) contain five rodent trapping webs. Each web consists of twelve 100m transects radiating out from a central rebar stake marked #145. There are four permanently marked ANPP plots on each of the trapping webs. These plots are located approximately 10 meters from the end of the transects extending in the four cardinal directions. Each plot consists of four 1m2 quadrats located around a tall center stake. Each quadrat is marked by two short fiberglass stakes. Quadrat 1 is northwest of the center stake with quadrats 2, 3, and 4 following in a clockwise direction. The grids are arranged in an 8-quad X 5-quad matrix. As of 2004, only quadrats 1 (northwest of center stake) and 3 (southeast of center stake) are sampled at each plot.

At the core_black site, webs 2 and 3 are no longer sampled because a natural fire burned these two webs. As a result, plots were added to the remaining webs (1, 4, and 5) to total 30 quadrats. On webs 1 and 5, a NE plot was added to the normal N, S, E, W total 5 plots. At web 4, the N plot was changed to NW and a NE plot was added, for a total of 5 plots (NW, NE, E, W, S).

Site core_PJ, the pinon-juniper woodland site at Cerro Montosa, is set-up differently than the other core sites. In order to accommodate the different habitat types, groups of transects (i.e., "plots") were established along north (N) and south (S) facing slopes as well as along vegas (V) and ridges (R). There are a total of 100 quadrats sampled at this core site. Transects on the first two plots consist of 40 quadrats each (10 quadrants for each of four habitat types). Plots 1 and 2 each have four transects: Vega (V), Ridge (R), South-facing slope (S), and North-facing slope (N). Each transect is made up of 10 quadrats, totaling 40 quadrats at each of these plots. Plot 3 is made up of four transects (A, B, C, D) that each contain 5 quadrats that capture the Piedmont habitat type, for a total of 20 quadrats. Plot 1 is north of Plots 2 and 3 and is accessible via a trail. Plots 2 and 3 are accessed by a trail leading to the weather station, Plot 3 is located just west of the weather station on the wide piedmont.

Note: Winter measurements of all sites except core_creosote ceased after 2006.

Note: On August 4, 2009, some of the webs and quadrats within the unburned Black Grama (core_black) site had a lightning-initiated fire. Thus, webs 2 and 3 were abandoned and extra plots added to areas within webs 1, 4, and 5 that were not burned. Changes were as follows: Webs 1, 4, and 5: A plot was added to the northeast to compensate for the loss of all plots at webs 2 and 3. Web 4: A plot was added to the northwest to compensate for the northern plot, which was burned.

Note: At the core_blue site, webs four and five are oblong rather than round. Therefore, the west and east plots are only 100 m apart.

Maintenance Core Sites data: 01/12/2010 - Data were QA/QC'd in a Navicat/mySQL pipeline until 2016. Metadata were updated and compiled for 1999-2010. (JMM) 11/29/2009 - Data were QA/QC'd and put in Navicat. Metadata were updated and compiled for 1999-2009. Note: In fall of 2009, data from site core_black, webs 2, 3, and 4 (plot N) were not collected due to unexpected fire at Sevilleta LTER sites. (YX) 01/05/2009 - Metadata were updated and compiled for 1999 - 2008. (YX) 01/06/2009 - As of 2007, winter season was no longer measured except at site core_creosote. (YX) 12/05/2009 - data from 1999-2008 were QA/QC'd in MySQL. 2006 (krw). In 2003, site core_blue was added. In 2004, the number of quadrats was reduced to 40 per site (quadrats 2 and 4 at each plot are no longer sampled). I checked for duplicates and missing quadrats. These most often happened when a recorder mislabeled a particular quad. I also checked every plant code against the USDA Plants database online at http://plants.usda.gov/. All plant codes that had nomenclature changes were updated. All previously unknown plants that have since been identified were also updated. All unknown plants that will never be identified were left in the database. All types were corrected. A list of codes not in the USDA list that are still in the data are as follows NONE = no plants in quadrat, OPUN = opuntia seedlings, SPOR = lumped Sporobolus spp (SPAI, SPCO4, SPCR, SPFL2), STEM = bare stem measurements for LATR2, U2 and UKFO18 and UKFO57 = unknowns that will never be identified, UKFO80 = unknown that has not yet been identified. A list of updates and the reason for the changes are below along with comments where identification is uncertain:

OLD CODE,NEW CODE,NUMBER_ROWS_AFFECTED,REASON_FOR_CHANGEPOOL,POOL,2,TYPO-999,BOER4,3,ERROR_IN_DATA_MANAGEMENT ALLI1,ALMA4,2,IDENTIFIED_UKFO AMAR2,AMPA,8,IDENTIFIED_UKFO AMAR3,ACNE,9,IDENTIFIED_UKFO AMAR4,AMPA,4,IDENTIFIED_UKFO APIA1,CYMO,2,IDENTIFIED_UKFO ARDR4,ARLUL2,45,BELIEVED_MISIDENTIFICATION ARLUA,ARLUL2,3,BELIEVED_MISIDENTIFICATION ASTE13,SCMU6,31,IDENTIFIED_UKFO ASTE5,UKSH5,4,STILL_UNKNOWN ASTE7,TOAN,1,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION ASTRA,ASMIM,1,BEST_GUESS_FROM_DESCRIPTION BRAS1,LEDED,1,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION BRAS2,DRGL5,54,IDENTIFIED_UKFOBR BR2,BRCA3,4,ONLY_CHANGE_TO_BRCA3_AFTER_NEW_PJ_PLOT_DESIGN BREU,BREUC2,1,TYPO BRIC1,BRBR2,1,IDENTIFIED_UKFO BRIC3,BREUC2,5,IDENTIFIED_UKFO BRIC4,BREUC2,1,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION BRIC5,UKSH5,1,"KNOWN_FROM_LOCATION,_STILL_UNKNOWN" CACT,OPUN,4,"IN_ROW_ID_13908,_13919,_AND_184 47_CHANGE_COVER_TO_1,_THESE_ARE_LENGTH_BY_WIDTH_MEASUREMENTS,_OPUTIA_SPP_SEEDLINGS" CACT1,CACT1,0,NEVER_TO_BE_IDENTIFIED CADR6,HODR,630,NAME_CHANGE CAJA6,POJA5,8,NAME_CHANGE CHAL2,CHAL11,2,CODE_REDUNDANCY CHAM,CHMI7,1,IDENTIFIED_UKFO CHCO2,CHCO,5,ONLY_AT_SITE_MS CHEN1,TECO,58,IDENTIFIED_UKFO CHGO2,CHCO2,1,TYPO CHLA2,CHLA10,127,CODE_REDUNDANCY COAR4,VINE,10,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" COAU,COAU2,1,TYPO COEQ,VINE,3,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" CONV1,VINE,3,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" CONV3,VINE,7,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" CONV4,VINE,6,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" CRAB,PAOB,2,IDENTIFIED_UKFO CYMO,HYFIC,4,ONLY_AT_SITE_B DAJA,DABR,7,ONLY_AT_SITE_P DEOBO,DEPI,675,BELIEVED_MISIDENTIFICATION DEWO?,DEWO,1,CONFIRMED_ID ECFEF,ECCOC,2,BELIEVED_MISIDENTIFICATION ECFEF2,OPUN,1,ONLY_AT_PIS4 ECFEF2,ECCOC,3,ONLY_AT_SITE_P ECFEF2,ECFEF3,2,NAME_CHANGE ERCI,ERCI6,11,ONLY_ON_5/26/04_SITE_P ERCI6,ERCI,21,ALL_SITE_B ERDI2,ERFL,17,BELIEVED_MISIDENTIFICATION ERDI4,ERFL,32,BELIEVED_MISIDENTIFICATION ERRO2,ERPO4,1,ONLY_AT_SITE_P ERPU8,DAPU7,2006,NAME_CHANGE SCIND,ESVIV,3,ONLY_AT_MG EUGL3,CHGL13,1,NAME_CHANGE FABA1,LUBR2,1,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION FABA3,LOPL2,4,IDENTIFIED_UKFOF ORB1,FORB1,0,STILL_UNKNOWN FORB3,DIWI2,3,IDENTIFIED_UKFO GARR1,GACO5,2,IDENTIFIED_UKFO HEOB,HENA,9,BELIEVED_MISIDENTIFICATION HIJA,PLJA,1350,NAME_CHANGE HOGL2,HODR,985,BELIEVED_MISIDENTIFICATION HYVE,MILI3,1,ONLY_AT_SITE_P IPCO2,VINE,116,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" IPCO3,VINE,5,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" IPLE,VINE,1,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" IPLO,IPLO2,3,TYPO JF1,GACO5,2,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION JF3,PLPA2,13,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION JF5,POOL,1,MOST_COMMON_PORTULACA JG1,ARPUP6,2,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION JG2,BOGR2,4,"GRASS_SEEDLINGS,_LIKELY_BOGR2" JUM0,JUMO,1,SPELLED_WITH_A_ZERO KRLA,KRLA2,5,TYPO_NO_KRLA_AT_SITE_MS LARER,LAOCO,204,BELIEVED_MISIDENTIFICATION LITH1,LIIN2,3,IDENTIFIED_UKFO MAGR10,MAPIP,24,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" MIMU,MIOX,5,ONLY_FOR_P2R6 MUMI,MUTO2,1,QUESTIONABLE MUSQ,MOSQ,97,CODE_REDUNDANCY NEIN,ECIN2,12,NAME_CHANGE NYCT1,BOSP,1,IDENTIFIED_UKFO NYCT2,MILI3,14,IDENTIFIED_UKFO OEAL,OECAC2,17,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" OECEC2,OECAC2,806,CODE_REDUNDANCY ONAG1,GASUN,16,IDENTIFIED_UKFO OPEN,OPEN3,31,TYPO OPMAC,OPMA8,10,TYPO OPUN,OPUN,3,"IN_ROW_ID_19570,_20831,_AND_21055_CHANGE_COVER_TO_1,_THESE_ARE_LENGTH_BY_WIDTH_MEASUREMENTS,_OPUTIA_SPP_SEEDLINGS" OPUN1,OPUN,1,"IN_ROW_ID_38109,_THIS_IS_A_SEEDLING,_CHANGE_COVER_TO_1" PEPA20,SCPA10,1,NAME_CHANGE PF1,DRCUC,24,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION PF2,ARLUL2,15,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION PF3,PF3,0,NEVER_TO_BE_IDENTIFIED PF4,PF4,0,NEVER_TO_BE_IDENTIFIED PG1,HENE5,8,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION PHHEF,SOJA,3,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" POAC1,POAC1,0,NEVER_TO_BE_IDENTIFIED POAC11,POFE,25,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION POAC12,PAOB,3,IDENTIFIED_UKFO POAC14,POAC14,0,NEVER_TO_BE_IDENTIFIED POAC7,LYPH,3,IDENTIFIED_UKFO POLY1,CHGR2,98,IDENTIFIED_UKFO PORT1,POOL,1,MOST_COMMON_PORTULACA QUGR3,QUTU2,1244,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" SAKA,SATR12,588,BELIEVED_MISIDENTIFICATION SC?,SCPA10,1,GRAMA_CACTUS SCIND,ECIN2,19,NAME_CHANGE SCINI,ECIN2,46,BELIEVED_MISIDENTIFICATION SCSCN2,BOCU,8,BELIEVED_MISIDENTIFICATION SEED2,BELY,6,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION SOLA6,SOJA,14,IDENTIFIED_UKFO SOLA7,SOJA,2,IDENTIFIED_UKFO_PHHEF_LUMPED_WITH_SOJA SPAI,SPOR,39,ONLY_IN_SITE_P SPCO4,SPOR,2288,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" SPCR,SPOR,3603,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" SPFL2,SPOR,2485,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" SPHAE,SPWR,5,MOST_LIKELY_SPHAERALCEA SPORO,SPOR,2,"BAD_IDENTIFICATIONS,_I_RECOMMEND_LUMPING" STNE,HENE5,6,NAME_CHANGE STNE2,HENE5,72,NAME_CHANGE U1,MAFE,3,IDENTIFIED_UKFO U2,U2,0,NEVER_TO_BE_IDENTIFIED U3,U3,0,NEVER_TO_BE_IDENTIFIED U4,U4,0,NEVER_TO_BE_IDENTIFIED U5,VUOC,26,IDENTIFIED_UKFO U7,SCLA6,3,IDENTIFIED_UKFO UKAS2,ERFL,7,BEST_GUESS_FROM_DESCRIPTION UKCA,CHFE3,2,MOST_COMMON_CHAEMACYSE_IN_AREA UKCA1,MAHEH2,1,LOOKED_IN_FUTURE__DATA UKFO,SEDI3,1,BEST_GUESS_FROM_DESCRIPTION UKFO10,UKFO10,0,NEVER_TO_BE_IDENTIFIED UKFO13,UKFO13,0,NEVER_TO_BE_IDENTIFIED UKFO15,ARLUL2,2,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION UKFO16,BRBR2,3,IDENTIFIED_UKFO UKFO17,UKFO17,0,NEVER_TO_BE_IDENTIFIED UKFO18,UKFO18,0,NEVER_TO_BE_IDENTIFIED UKFO19,GUSA2,3,IDENTIFIED_UKFO UKFO20,SPLE,2,BEST_GUESS_FROM_DESCRIPTION UKFO21,DAJA,1,IDENTIFIED_UKFO UKFO22,SEDI3,1,IDENTIFIED_UKFO UKFO23,MESCS,1,IDENTIFIED_UKFO UKFO31,UKFO31,0,"COULD_BE_BADI,_GLWR_OR_NECA3" UKFO32,SACYH2,1,IDENTIFIED_UKFO UKFO51,UKFO51,0,NEVER_TO_BE_IDENTIFIED UKFO57,UKFO57,0,NEVER_TO_BE_IDENTIFIED UKFO61,THWR,186,IDENTIFIED_UKFO UKFO62,THWR,34,IDENTIFIED_UKFO UKFO7,ZIGR,2,IDENTIFIED_UKFO UKFO72,MILI3,27,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION UKFO72?,MILI3,1,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION UKFO73,HYVE,4,PROBABLE_IDENTIFICATION_FROM_UKFO_DESCRIPTION UKFO75,UKFO75,0,NEVER_TO_BE_IDENTIFIED UKFO76,UKFO75,3,NEVER_TO_BE_IDENTIFIED UKFO80,UKFO80,0,NOT_YET_IDENTIFIED UKGR2,LYPH,1,BEST_GUESS_FROM_DESCRIPTION UKSH1,SPLE,2,BEST_GUESS_FROM_DESCRIPTION UKSH4,BREUC2,8,IDENTIFIED_UKFO UKSH5,UKSH5,0,NOT_YET_IDENTIFIED

GRIDS -

Begun in spring 2013, this project was part of a long-term study at the Sevilleta LTER measuring net primary production across three distinct ecosystems: creosote-dominant shrubland (grid_creosote), black grama-dominant grassland (grid_black), and blue grama-dominant grassland (grid_blue).

Sampling Quadrats: Each sampling grid contains 40 1x1m quadrats in a 5x8 array. However, only 30 quadrats are sampled at each. These are quadrats 1-15 and 26-40. Thus, the middle two rows (i.e., 10 quadrats) are not sampled.

Locating the Sampling Quadrats: Three core sites (B, G, and C) contain five rodent trapping and vegetation sampling webs. The vegetation grids are near these webs at each core site. At the blue grama site, the grid is located at the southern end of web 5, between webs 2 and 4. At the creosote site, the grid is east of web 3, near the road. At the black grama site, the grid is just northeast of web 5.

BURN -

In 2003, the U.S. Fish and Wildlife Service conducted a prescribed burn over a large part of the northeastern corner of the Sevilleta National Wildlife Refuge. Following this burn, we designed a sampling scheme to detect effects of fire for three different vegetation types: mixed grass, mixed shrub and grassland, dominated by black grama grass (Bouteloua eriopoda).

Forty permanent 1m x 1m plots were installed in both burned and unburned (i.e., control) sections of each habitat type. The core_black grama site is used as an unburned control site for analyses. The mixed_grass control site caught fire unexpectedly in the fall of 2009 and some plots were subsequently moved to the south. For details of how the fire affected plot placement, see below. In spring 2010, sampling of plots 16-25 was discontinued at the mixed_grass (burned and control) and grassland_burn (burned treatment only) sites, reducing the number of sampled plots to 30 at each site.

Burn: On August 4, 2009, a lightning-initiated fire began on the Sevilleta National Wildlife Refuge. The fire reached the mixed_grass unburned plots on August 5, 2009, consuming them entirely. As a result, in the spring of 2010, the mixed_grass unburned plots were moved to a different area within Deep Well, southwest of the Warming site. Also, on August 4, 2009, some of the webs and quadrats within the unburned Black Grama (core_black) site were burned by the fire. Thus, webs 2 and 3 were abandoned and extra plots added to areas within webs 1, 4, and 5 that were not burned. Changes were as follows: Webs 1, 4, and 5: A plot was added to the northeast to compensate for the loss of all plots at webs 2 and 3. Web 4: A plot was added to the northwest to compensate for the northern plot, which was burned.

Maintenance: Data QA/QC 01/13/2011-Burn NPP quad data was QA/QC'd and put in Navicat. Matadata updated and compiled from 2004-2010. The mixed-grass unburned plot was moved to the south after the original plot burned unexpectedly in the fire of August 2009. (JMM) 11/28/2009-Burn NPP quad data was QA/QC'd and put in Navicat. Metadata updated and compiled from 2004-2009. Mixed-grass unburned data (Fall 2009) was not collected due to unexpected fire at Sevilleta LTER in Aug 2009. (YX) 01/14/09-Metadata updated and compiled from 2004-2008 data. As of 2007, winter measurements are longer being taken. (YX) 12/20/2008-This data was QAQC'd in MySQL. I checked for duplicates and missing quadrats. (YX)

CRUST -

The loss of biological soil crusts (biocrusts) is occurring in dryland ecosystems worldwide. Causes include land use intensification, grazing, human disturbance of soils (trails, ATVs), and climate change. Consequences of the loss of biocrusts include altered soil surface hydrology, changes to soil C and N dynamics, and increased soil erosion and dust production. Much less is known about the changes in community composition that may accompany loss of biocrusts. This project experimentally tests how soil disturbances that reduce biocrusts affect the composition of SEV communities, including plants. The work is directly related to the core activities of the Sevilleta LTER to understand the ecological dynamics of disturbances to dryland ecosystems. Information will help to interpret the potential community and ecosystem-level impacts of on-going climate manipulations on biocrusts via our direct manipulation of biocrust abundance. Our experiments have potential to expand current knowledge on the biological importance of biocrusts for dryland communities and ecosystems.

This project is located in three ecosystem types of the Sevilleta National Wildlife Refuge: a mixed creosote shrubland-black grama grassland site (crust_creosote), a mixed blue grama-black grama grassland (crust_grass), and grassy areas within pinon-juniper woodland (crust_PJ, added May 2018). At each site, we identified 20 plots (6 m X 6 m) with similar initial plant and biocrust cover and plant species composition (10 disturbed plots and 10 controls per site). Each site contains 20 plots, with two quadrats per plot, for a total of 120 quadrats. Plots were arranged in a 4 5 grid in creosote and grass, but along transects at PJ to maximize the presence of patchily distributed biocrust. We imposed a minimum distance of ~20 m between plots, and we continually minimize foot traffic at the site by walking between plots on fixed pathways that parallel the eastern and southern borders of each plot. Pre-treatment data collection began May 2013 at crust_grass and crust_creosote sites and May 2018 at PJ.

Soil disturbance treatments: To determine potential impacts of the loss of biocrusts, we imposed soil disturbance that reduced biocrust abundance. Soil disturbance occurs twice yearly, after data collection in May and October, and involves disturbing all non-vegetated soil surfaces to a depth of ~ 5cm. Control plots are not disturbed. Disturbance involves two personnel wearing thickly treaded boots, each stepping on and scuffing up all non-vegetated soil surfaces in the plot. Using this method, each non-vegetated patch in the plot receives two "stomps," and stomping disturbs the soil surface to a depth of ~5cm. The disturbance treatment was assigned to plots in a completely randomized design, and beginning June 2013 (May 2018 for PJ), disturbance was imposed twice each year (May, Oct) to capture the beginning and end of the growing/monsoon season.

Cover and height estimates are taken for each plant species within each of two permanently marked 1 m X 1m quadrats per plot. Quadrat corners are permanently marked with large nails inserted into the soil. Pre-treatment data were taken in four quadrats per plot (May 2013), and repeated measures data (Sept 2013 onwards) are taken in a subset of two of the four quadrats, usually quadrats 1 and 3, unless a disturbance (e.g., kangaroo rats, seed harvester ants) necessitated a different quadrat. In plot 29, quadrats 1 and 4 are sampled

EDGE -

EDGE is located at six grassland sites that encompass a range of ecosystems in the Central US - from desert grasslands to short-, mixed-, and tallgrass prairie. EDGE is a research platform to advance our understanding of patterns and mechanisms of ecosystem sensitivity to climate change, but it also benefits the broader scientific community by creating room for additional experiments and observations. Identical infrastructure for manipulating growing season precipitation was deployed at all sites. Treatment plots are 6m x 6m.

Study Sites. The six sites were selected to capture the key environmental and ecological gradients of Central US grasslands and represent the major grassland ecosystem types (desert, shortgrass, mixedgrass, and tallgrass) of the region. Site selection criteria included: site characteristics (mean annual precipitation and temperature, dominant vegetation), access and site security, permission to build experimental infrastructure, participation in an existing or future network (e.g., LTER, NEON), and available site support and supporting data (e.g., LTER, USFWS or ARS).

Experimental Treatments and Plots. We imposed a significant reduction in growing season precipitation (-66 % of ambient) during 2013-2019. This is the equivalent of a ca. 50% reduction in annual precipitation because at all sites about 60-75% of annual precipitation falls in the growing season. We imposed this long-term drought either by reducing the size of each rainfall event (event size reduction, E, -66 % of ambient) or by delaying the onset of the monsoon season rains (delay, D). The control (C) treatment is included for comparison. For the event size reduction treatment (E), each rainfall event was passively reduced by a fixed proportion. Note that rain event number and the average number of days between events does not differ from ambient treatment. For the delay (D) treatment, shelters roofs are 100% exclusions that catch all rain into large rain barrels. Delay shelters are in place from mid-July to mid-September, then all rain caught during that period is re-applied to plots during September-October.

Plot Setup. At each site, we established replicate 6 m x 6 m experimental plots (n = 10 per treatment, including the control treatment) in a relatively homogeneous area (similar soils, vegetation, etc.) that is representative of the overall site. Plots are arrayed such that each treatment is co-located in a single block (n=10 blocks per site), with each block located at least 5 m apart. The blocking controls for environmental gradients and standing genotype X distance effects (Whitney et al. 2019 Oecologia 10.1007/s00442-019-04371-7), if present. For each site, all plots within a block (including the control) are located at least 2 m apart and trenched to 20-30 cm depth with aluminum flashing to hydrologically isolate them from the adjacent soil, and each plot is covered by the rainfall manipulation infrastructure. The 6 m x 6 m plot size includes a 0.5 m external buffer to allow access to the plots and minimize edge effects associated with the infrastructure. The resulting 5 x 5 m area was divided into 4 2.5 x 2.5 m subplots. Two subplots were designated for plant species composition sampling. Four two quadrats (quadrats 2 and 3) were sampled in each plot, for a total of 120 quadrats per site and 240 quadrats total. Each plot has a pole with a metal tag showing the plot number. The quadrat closest to the pole is quad 1, and immediately next to it is quad 2. Quad 3 is at the corner of the plot diagonal from quad 1, and quad 4 is immediately next to quad 3. As of fall 2019, only quadrats 2 and 3 are sampled

Rainfall Manipulation Infrastructure We passively altered rainfall reaching the plots by using a version of a rainfall reduction shelter designed by Yahdjian and Sala (2002). The most significant environmental artifacts of these shelters are a 5-10% reduction in light due to the acrylic V-shaped shingles and a ~ 20 cm edge effect (Yahdjian and Sala 2002). Shelters consist of a steel frame that supports a roof.

FERTILIZER -

This long-term study at the Sevilleta LTER, began in spring 2004 to examine how nitrogen fertilization affects plant species in a mixed desert-grassland. Plant species composition and the cover and height of individuals, are sampled twice yearly (spring and fall) at permanent 1m x 1m plots.

Study Design: Twenty 10m X 5m plots were established in December 2005. Ten plots are fertilized; ten are not. Fertilizer was applied twice yearly (spring and fall) as granular NH4NO3 at the rate of 100 kg N ha-1 yr-1 until 2019, at which point we switched to yearly fertilization during July, prior to the onset of the monsoon season. Within each plot are four sampled quadrats (n=80 quadrats). Quadrats are adjacent and numbered from south (1) to north (4) starting at the southeastern corner of each plot. All quadrats were sampled until 2019, at which point we analyzed data and reduced quadrats to sampling only quadrats 3 and 4.

1/12/2011 - Quad data for spring and fall 2010 QA/QC'd and put in Navicat. Metadata updated and compiled for 2010. (JM) 11/28/2009 - Quad data QA/QC'd and put in Navicat. Metadata updated and compiled for 2006 -2009. (YX) 1/06/09 - Metadata updated and compiled for 2006, 2007, 2008. (YX) 01/06/09 - As of 2007, winter measurement is no longer being taken. (YX) 12/10/2008 - NPP data from 2004-2008 was QAQC'd in MySQL. I checked for duplicates and missing quadrats. (YX) 1/12/2006 - Meta data compiled for 2005. 5/16/2005 - Season 3, Plot 11 data didn't distinguish between quadrats 3 and 4. Original data were checked and this error was corrected. -KLV

ISOWEB -

Monthly mammal trapping, including stable isotope analysis, is carried out at two webs (iso_web) on the SEV, located on either side of McKenzie South Road. Since 2018, vegetation is sampled on these webs to allow for analysis of relationships between mammal populations and plant communities.

There are two webs, with four plots per web and 4 quadrats per plot. Web 1 is southeast of web 2. At web 1, the plots are located at the ends of transects 1, 4, 7 and 10; at web 2, the plots are at the ends of transects 3, 5, 7 and 10. Plots are marked with rebar stakes. The plot closest to the road is plot 1, and plot numbers ascend as you go clockwise. As with the core webs, quad 1 is the northwest quad of each plot , and the quad numbers ascend as you go clockwise; quad 1 is tagged. Unlike at the core webs, all four quadrats in each plot are sampled, for a total of 32 quadrats.

MEAN-VARIANCE -

The Mean Variance experiment is designed to investigate the effects of changes in precipitation mean and precipitation variance. Forecasting the consequences of climate change is arguably the most pressing challenge at the interface of science and society. Not only is mean temperature increasing, but precipitation is becoming more variable. Most prior ecological research on climate change has emphasized trends in mean climate variables or separate study of extreme events. Yet, effective forecasts require determining responses to both non-stationary components of climate distributions: the mean and the variance. Effects of variance arise from nonlinearities in ecological responses and stochasticity in climate events, and confronting variance has dramatically transformed some ecological disciplines. However, empirical resolution of variance effects has lagged behind theory because these effects play out over timescales that exceed standard funding cycles, making long-term support critical to this scientific frontier. NSF-LTER funding enabled the first experiment to combine a drier mean climate with increasing interannual variance in a factorial design. This signature experiment was built in Plains grassland ecosystem in summer 2019, and the Chihuahuan Desert grassland in summer 2020. The experiment has three novel elements: (i) determining interactive effects by crossing reduced mean with increased variance, (ii) replicating in multiple dryland ecosystems to compare their susceptibilities, and (iii) experimentally adding stochasticity to permit the antecedent effects that occur under natural climate variability.

The meanvar_blue site is located at the blue grama core site, between the Blue grama met station and the road. There are 30 plots arranged in blocks of three in a snaking pattern, with two quadrats per plot (quad 1 to the north, quad 2 to the south) for a total of 60 quadrats.

Precipitation treatments: To increase variance in soil moisture (θ) stochastically, without changing the mean, we paired plots and amplified their precipitation regimes. Every water year, plots within a pair are randomly assigned to either a 50% decrease or 50% increase in precipitation. We create stochasticity through random assignments of plots to treatments in each year. To reduce the long-run mean soil moisture, we intercept 25% of precipitation, a moderate forcing within range of likely futures. All plots are covered with tented roof panels made of acrylic plastic V-shaped shingles that intercept precipitation year-round. Control plots have shelters but with panels that allow precipitation through by flipping the V shape of each shingle. Plots receiving both reduced mean and increased variance will randomly alternate between 75% less net precipitation (-25% for mean - 50% for variance) or 25% more (-25% for mean + 50% for variance). Water is captured from shelters with gutters, stored in tanks (black plastic rain barrels), then delivered to the paired plots via a solar-powered pump. Replication is uneven to account for higher variability among plots in the increased variance treatment (N - 5 for mean treatments, and 10 for variance treatments). All plots are hydrologically isolated via aluminum flashing installed to 20-30 cm depth, with a 5 cm ridge above the soil surface.

Plot size: 5m x 5m

Plot height ~3 m

Number of plots per site: 30

Number of tanks per site: 10

Sprinklers: 10 plots have sprinklers to deliver rainwater captured from the rainfall reduction plots

MRME -

Begun in fall 2006, this long-term study at the Sevilleta LTER examines changes in plant community structure and primary production caused by increased intra-annual variability in monsoon rainfall. Increased precipitation variability within a season alters the pulses of soil moisture that drive primary productivity, community composition, and ecosystem functioning. The overarching hypothesis is that changes in event size and variability will alter grassland productivity, ecosystem processes, and plant community dynamics. We predict that many small events will increase soil CO2 effluxes by stimulating microbial processes but not plant growth, whereas a small number of large events will increase aboveground net primary production and soil respiration by providing sufficient deep soil moisture to sustain plant growth for longer periods of time during the summer monsoon.

Experimental Design: The Monsoon Rainfall Manipulation Experiment (MRME) contains three ambient precipitation plots and five replicates of the following treatments: 1) ambient plus a weekly addition of 5 mm rainfall; 2) ambient plus a monthly addition of 20 mm rainfall. Rainfall is added during the monsoon season (July-Sept) by an overhead (7 m) system fitted with sprinkler heads that produces rainfall-quality droplets. At the end of the summer, each treatment has received the same total amount of added precipitation, delivered in different sized events. Each plot (9 m x 14 m) includes subplots (2 m x2 m) that receive 50 kg N ha-1 y-1. Measurements include: (1) seasonal (July, August, September, and October through June) soil N; (2) plant species composition and annual net primary production (ANPP); (3) seasonal root and fungal dynamics within mini-rhizotrons, and; (4) soil temperature, moisture, and CO2 fluxes (using in situ solid state CO2 sensors). In addition, soil N2O fluxes, predawn and mid-day (10:00-12:00) water potential, and the mid-day leaf photosynthetic gas exchange and stomatal conductance of black grama are measured prior to and up to five days after scheduled precipitation events. Derivation of Biomass and NPP: Data from SEV188 and SEV157 are used to calculate seasonal and annual production of each species in each quadrat for a given year. Allometric equations derived from harvested samples of each species for each season are applied to the measured cover, height, and count of each species in each quadrat. This provides seasonal biomass for winter, spring, and fall. Seasonal NPP is derived by subtracting the previous season's biomass from the biomass for the current season. For example, spring NPP is calculated by subtracting the winter weight from the spring weight for each species in a given quadrat. Negative differences are considered to be 0. Likewise, fall production is computed by subtracting spring biomass from fall biomass. Annual biomass is taken as the sum of spring and fall NPP. August 2009 Burn: On August 4, 2009, a lightning-initiated fire began on the Sevilleta National Wildlife Refuge. The Monsoon site was entirely burned on this date, with all plots subjected to fire of comparable intensity.

The Monsoon Rainfall Manipulation Experiment (MRME), located NW of Five Points past EDGE Black on Seca Road, examines changes in ecosystem structure and function of a semiarid grassland caused by increased precipitation variability, which alters the pulses of soil moisture that drives primary productivity, community composition, and ecosystem functioning.

This study has 13 plots that each receive one of three treatments. Five plots receive small watering treatments, five plots receive large watering treatments, and three plots serve as controls and receive no additional water treatment. Each plot contains two adjacent subplots, subplots A and B. One subplot receives a fertilizer treatment and one does not. Each subplot is divided into four quadrats. Only quadrats 1 and 2 are sampled. Quadrat 1 is located in the northwest corner of the subplot while quadrat 2 is diagonal from quad 1, in the southeast corner of the subplot. There is a total of 52 quadrats sampled at this site.

Maintenance: File created 3/2/2005. -- KLV Updated 12/11/2006 --KRW Data appended to file on 7/25/2005 -- KLV Data compiled into one file. Metadata entered in EML access database. TK 6 February 2009 data qa/qc in navicat. Made NONE measurements in the following format Cover 0 Height -888 Count -888, which have since been converted to NA. Corrected typos and errors. TLK 10 February 2009

NutNet -

This experiment seeks to determine how nutrient availability controls plant biomass, diversity, and species composition in a desert grassland. Two of the most pervasive human impacts on ecosystems are alteration of global nutrient budgets and changes in the abundance and identity of consumers. Fossil fuel combustion and agricultural fertilization have doubled and quintupled, respectively, global pools of nitrogen and phosphorus relative to pre-industrial levels. In spite of the global impacts of these human activities, there have been no globally coordinated experiments to quantify the general impacts on ecological systems. This has important implications for understanding how future atmospheric deposition of nutrients (N, S, Ca, K) might affect community and ecosystem-level responses. This study is part of a larger coordinated research network that includes more than 40 grassland sites around the world. By using a standardized experimental setup that is consistent across all study sites, we are addressing the questions of whether diversity and productivity are co-limited by multiple nutrients and if so, whether these trends are predictable on a global scale.

The SEV Nutrient network (NutNet) site, located just west of the warming plots near Deep Well, is part of a larger coordinated research network that includes more than 40 grassland sites around the world. This experiment seeks to determine how nutrient availability controls plant biomass, diversity, and species composition. The Sev SEV experiment was initiated in 2007. The experimental design is completely randomized with 8 treatments replicated 5 times each. The nutrients added include N (nitrogen), P (phosphorus), and K (potassium plus other nutrients). Treatments are: +N+P+K, +N+P, +N+K, +N, +P+K, +P, +K, and control (no nutrients added), administered once a year in the summer months . Treatments were randomly assigned to 40 plots (5 m X 5 m) with 1m of buffer separation between every plot. The treatments are marked on stakes in each plot. Each plot contains one quadrat, marked at the corners with fiberglass stakes.

Methods: Nutrient addition treatments and sampling sites are located in an area of desert grassland dominated by black grama, Bouteloua eriopoda. The experimental design is completely randomized with 8 treatments replicated 5 times each. The nutrients added include N (nitrogen), P (phosphorus), and K (potassium plus other nutrients). Treatments are: +N+P+K, +N+P, +N+K, +N, +P+K, +P, +K, and control (no nutrients added). Treatments were randomly assigned to 40-25 m2 plots with 1m separating each plot. Response variables measured include: plant community composition; percent ground cover of live perennial grasses, herbaceous dicots, shrubs, cactus, litter, and bare ground; aboveground net primary production; light availability, and several soil parameters (moisture, organic matter content, pH, P, field available nitrogen (NO3-N and NH4-N), potentially mineralizable N). This experiment was initiated in May 2007 with one year of pre-treatment data and 3 years of post-treatment data collected thus far. Nutrients are applied annually at the beginning of the growing season starting in 2008. After 2017, nutrient applications were always done in July.

The following dates are NutNet collections, some are poorer quality data, please indicate which ones. 5/13/2008, 9/19/2008, 5/1/2009, 5/26/2010, 10/14/2010, 5/25/2011, 11/12/2011, 5/5/2012, 10/1/2012, 5/3/2013, 9/1/2013, 9/2/2013, 4/22/2014, 4/24/2014, 9/16/2014, 9/23/2014, 4/23/2015, 10/13/2015, 4/18/2016, 9/28/2016, 10/10/2017, 5/3/2018, 10/8/2018

Additional Information. Species composition and net primary production was sampled semiannually (spring and fall) in 2007, 2008, and 2009. Soil was sampled and analyzed in the fall in 2007 and 2008. Plots were fertilized annually starting in 2008. In August 2009, a wildfire burned all 40 of the NutNet plots causing no Fall 2009 vegetation measurements. Special Codes for Vegetation Ids: SPORSP- Unknown Sporobolus SPSP- Unknown Sphaeralcea UNKFO- Unknown Forb On 08/20/2015, the following taxonomic changes were made to the data: ARPUP6 was changed to ARPU9, OECAC2 was changed to OECA10, SPWR was changed to SPPO6.

TOWER -

The varied topography and large elevation gradients that characterize the arid and semi-arid Southwest create a wide range of climatic conditions - and associated biomes - within relatively short distances. This creates an ideal experimental system in which to study the effects of climate on ecosystems. Such studies are critical given that the Southwestern U.S. has already experienced changes in climate that have altered precipitation patterns and stands to experience dramatic climate change in the coming decades. Climate models currently predict an imminent transition to a warmer, more arid climate in the Southwest US. Thus, high elevation ecosystems, which currently experience relatively cool and mesic climates, will likely resemble their lower elevation counterparts, which experience a hotter and drier climate. In order to predict regional changes in carbon storage, hydrologic partitioning and water resources in response to these potential shifts, it is critical to understand how both temperature and soil moisture affect processes such as evapotranspiration (ET), total carbon uptake through gross primary production (GPP), ecosystem respiration (Reco), and net ecosystem exchange of carbon, water and energy across elevational gradients. We use six widespread biomes along an elevational gradient in New Mexico -- ranging from hot, arid ecosystems at low elevations to cool, mesic ecosystems at high elevation to test specific hypotheses related to how climatic controls over ecosystem processes change across this gradient.

We have an eddy covariance tower and associated meteorological instruments in each biome which we are using to directly measure the exchange of carbon, water and energy between the ecosystem and the atmosphere. This gradient offers us a unique opportunity to test the interactive effects of temperature and soil moisture on ecosystem processes, as temperature decreases and soil moisture increases markedly along the gradient and varies through time within sites. This dataset examines how different stages of burn affects above-ground biomass production (ANPP) in a mixed desert-grassland. Net primary production is a fundamental ecological variable that quantifies rates of carbon consumption and fixation.

Plant species cover and height is sampled at two flux tower sites, tower_east and tower_west, located near warming on either side of the road.. Each site has 20 quadrats for a total of 40 quadrats. They are aligned in an L-shape, with one leg radiating south from the tower and the other radiating east from the tower. TOWER quadrats are missing data during 2017 and fall 2018 due to lack of LTER funding that affected the number of staff.

WARMING (WENNDEx) -

A multi-factorial field experiment in an arid grassland within the Sevilleta National Wildlife Refuge (NWR) simulates increased nighttime temperature, higher N deposition, and heightened El Nino frequency (which increases winter precipitation by an average of 50%). Humans are creating significant global environmental change, including shifts in climate, increased nitrogen (N) deposition, and the facilitation of species invasions. The purpose of the experiment is to better understand the potential effects of environmental change on grassland community composition and the growth of introduced creosote seeds and seedlings. The focus is on the response of three dominant species, all of which are near their range margins and thus may be particularly susceptible to environmental change. It is hypothesized that warmer summer temperatures and increased evaporation will favor growth of black grama (Bouteloua eriopoda), a desert grass, but that increased winter precipitation and/or available nitrogen will favor the growth of blue grama (Bouteloua gracilis), a shortgrass prairie species. Furthermore, it is thought that the growth and survival of introduced creosote (Larrea tridentata) seeds and seedlings will be promoted by heightened winter precipitation, N addition, and warmer nighttime temperatures. Treatment effects on limiting resources (soil moisture, nitrogen mineralization), species growth (photosynthetic rates, creosote shoot elongation), species abundance, and net primary production (NPP) are all being measured to determine the interactive effects of key global change drivers on arid grassland plant community dynamics.

Burn: On August 4, 2009, a lightning-initiated fire began on the Sevilleta National Wildlife Refuge. By August 5, 2009, the fire had reached the Warming site, which was burned extensively though not entirely. Approximately 50% of plots burned on August 5 and those plots which did not burn were burned within three weeks by US Fish and Wildlife. Thus, the condition of all plots at the Warming site was comparable by early September 2009.

Experimental Design: This is a multi-factorial experiment with three fully crossed factors: warming, water addition, and nitrogen addition in a complete randomized design for a total of eight treatment combinations. There are five replicates for each of the eight treatment combinations and a total of 40 plots. Plots are 3 x 3.5 meters. This study comprises 40 plots; each plot is 3m X 3.5m. Each plot contains two quadrats marked with short PVC: Quadrat 1 is located to the north and Quadrat 2 is located to the south. The plot numbers are found on PVC piping next to the plot frame. There are 80 quadrats sampled at this site.

Warming System: The warming system consists of automated nighttime warming roofs made of aluminum fabric. The fabric reflects longwave radiation from the ground that would otherwise be lost to the atmosphere. This treatment is fully automated.

Water addition: Water is added using an overhead irrigation system. Water is trucked in, pumped from the road to storage tanks adjacent to the plots, and then pumped from the tanks to plots using a custom-designed overhead irrigation system that waters four plots at once. Sprinkler heads and orientation were chosen to mimic natural rain droplet size and velocity. From January through March there are 4x5mm applications, 1x10mm application and 1x20mm application. Exact timing of applications can vary based on weather and the necessary avoidance of temperatures below freezing and high wind.

Nitrogen addition: Nitrogen is added at the rate of 2 g/m2/yr in two equal aliquots. Thus, 1 g/m2/yr is added in February and 1 g/m2/yr in early July, following the beginning of the monsoon. Nitrogen is added in the form of ammonium nitrate (NH4NO3) pellets. 2.86 grams of ammonium nitrate yields 1 gram of nitrogen. Thus, each application puts 30 grams of nitrogen over each plot. In an effort to get an even distribution, the nitrogen is weighed out in several (4-6) portions and hand broadcast over the plot as evenly as possible.

Maintenance: 01/14/11-Spring and fall 2010 data was QA/QC'd and entered into Navicat. Metadata was updated and compiled for 2010. Winter 2010 data was not collected due to the fires of fall 2009. (JMM) 11/28/09-Quad data were QA/QC'd and put in Navicat. Metadata updated and compiled for 2006-2009. Fall 2009 data was not collected due to unexpected fire at Sevilleta NWR in Aug. 2009, and prescribed fire (Sep. 2009) at warming site. (YX) 01/06/09 Metadata created and compiled for 2006, 2007, 2008 data. (YX) 01/05/09 As of 2007, winter measurements are no longer being taken. I checked for duplicates and missing quadrats. (YX)