PROJECT TITLE: Nutrient Cycling in Tabonuco Forest (MRCE/LTER Litterfall Nutrient Fluxes, and Extractable Soil versus Microbial Biomass Nitrogen)

PROJECT DESCRIPTION: Understanding the long-term impact of deforestation on ecosystem structure and function of tropical forests may aid in designing future conservation programs to preserve biodiversity and sustain ecosystem productivity. We examined forest structure, tree species composition, litterfall (fine and coarse) due to Hurricane Hugo and subsequent fine annual litterfall inputs, litterfall rate, and leaf litter decomposition. The experiment was initiated by the MRCE (Minortity Research Centers of Excellence) program, and continued by the LTER. In addition to measuring nutrient fluxes from litterfall and decomposition, we measured KCl-extractable soil nitrogen in ammonium and nitrate forms to determine the effects of complete fertilization and removal of hurricane debris.

History of litterfall, decomposition and soil nitrogen availability studies (MRCE experiment) at the Luquillo Experimental Forest (LEF):

The Minority Research Centers of Excellence (MRCE) experiment was designed originally to determine whether forest productivity was limited by nutrient availability, genetic constraints or climatic variables along a steep environmental gradient in the Luquillo Mountains of Puerto Rico. Comparisons were made between dwarf forest at 500 m elev. and tabonuco forest from 300-400 m elev. There were two main experiments in the original design: 1) forest fertilization (complete versus none); 2) transplant experiments using common gardens at high and low elevation, with and without wind protection at high elevation. In addition, phenology of leaves was studied in the dwarf forest to determine the longevity of leaves in the canopy.

Hurricane Hugo struck in September 1989, following a year of pre-treatment measurements and just as the forest plot fertilization experiments were to begin, and deposited a year and a half of annual above ground litter inputs of phosphorus on the forest floor in green leaf litter (Lodge et al. 1991).  As phosphorus is thought to be the most limiting element, this necessitated the addition of a hurricane debris-removal treatment as a second type of control in the lower elevation tabonuco forest. Furthermore, the hurricane changed the nature of the study into one of looking at changes in forest composition and the recovery of forest productivity in response to fertilization and hurricane debris-removal. Seedling, sapling, herbaceous plant, and fern responses, as well as measurements of light availability and canopy closure were added to the originally planned measurements of tree diameter growth, leaf litter production, and fine root production and turnover. Funding for continued studies of the MRCE plots ended in 1998, but they were of such great value as a long-term experiment that they were incorporated into the LTER program. Another hurricane in 1998 (Hurricane Georges) presented an opportunity to look at the effects of removing only the woody debris (in new plots) versus removing all hurricane debris.

The leaf decomposition experiment was originally designed to examine short-term disappearance of foliage from three important montane species: Prestoea montana (R. Grah.) Nichols, Dacryodes excelsa (Vahl.), and Cyrilla racemiflora. Hurricane Hugo (August, 1989) provided a unique opportunity to study effects of this type of natural disturbance on decomposition. The study was repeated in 1990 as part of the LTER, with some changes in the design that allowed for comparisons of short -term foliar litter biomass and nitrogen dynamics: (1) among the three species, (between the colorado and tabonuco forest types, (3) between riparian and upland sites, and (4) between pre- and post-hurricane environments.

LTER CORE AREAS: (Annotate all that apply)

Primary Productivity

Organic Matter Accumulation

LEF LTER 1 RESEARCH TOPIC: (Annotate all that apply)

Disturbance regime

Recovery after disturbance

We define a data file as a component of a data set. A data set can have only one data file or more. Basically, different data files have different data structures or format.
DATA SET FILES (SUBSETS):

Data File No.	Data File Identifier	On-Line Filename	Starting Date	Periodicity of sample	End Period
1	Bisley Aboveground litterfall 1989-1992 data	abov8992.txt	August 28, 1989	every 2-4 weeks	May 18, 1992
2	Bisley Aboveground litterfall 1994-1995 data	abov9495.txt	June 13, 1994	every 2-4 weeks	June 6, 1995

See Web site at Yale for results in table format: belowground litter transfers along topographic gradients , aboveground litter transfers along topographic gradients

RESEARCH LOCATION: La Prieta, El Verde and Bisley #3, Bisley, Puerto Rico

INVESTIGATORS:

PRINCIPAL INVESTIGATORS E-MAIL address

Kristiina Vogt	kvogt@u.washington.edu
Daniel Vogt	dvogt@u.washington.edu
Tom Siccama	Thomas.Siccama@yale.edu
Alan Covich	alanc@cnr.colostate.edu
Fred Scatena	fns@sas.upenn.edu

OTHER RESEARCHERS E-MAIL address

Xiao-Yun Wang
Paul Boon	paul.boon@vu.edu.au
Janine Bloomfield
John Ranciato
Javier Perez

CONTACT PERSONS E-MAIL address Phone Number (Include area code)

Kristiina Vogt	kvogt@u.washington.edu	(203) 432-5076
Daniel Vogt	dvogt@u.washington.edu	(203) 432-5076

SOURCE OF FUNDING (SPONSOR): NSF Ecosystem Program, Andrew W. Mellon Foundation, and NSF-LTER

DATA SET ABSTRACT: Litterfall (fine and coarse) due to Hurricane Hugo and subsequent fine annual litterfall inputs (1, 2 and 5 yr after Hugo) were determined for two sites (El Verde and Bisley) in the Luquillo Experimental Forest in Puerto Rico. Litter transfers into streams, riparian and upslope areas were determined within each catchment. The recovery rate of aboveground fine litterfall (leaf, fine wood <1 cm diameter, and other miscellaneous inputs) to predisturbance levels were determined 1, 2, and 5 yr after Hurricane Hugo. The amount of total litter transfers and their individual components into the riparian and upslope areas due to Hurricane Hugo varied significantly by catchments within the Luquillo Experimental Forest. At El Verde, 26-39%, 31- 35%, 14-35% and 7-12% of the total litter transfers were contributed by leaf litter, fine wood, coarse wood and fine roots, respectively. At Bisley, 28-31%, 26-29%, 33-35% and 8-10% of the litter transfers were contributed by the same categories. Differential decay rates contributed to the relative importance of fine and coarse litter inputs. The recovery of fine aboveground litterfall to pre-hurricane levels after 5 yr varied by topographic location (streams had the slowest recovery, upslope areas the highest) and catchment (El Verde: 55-77%; Bisley: 39-82% of pre-hurricane values).

DATA SET METHODS: This study was conducted in a Tabonuco forest type classified as a subtropical wet forest in the Luquillo Experimental Forest in northeastern Puerto Rico. The dominant tree species was Dacroydes excelsa with Prestoea montana and Sloanea berteriana as commonly associated species (Brown et al. 1983). Soils were classified as clay and silty clay loam ultisols in the Los Guineos soil series (Boccheciamp 1977). Annual rainfall is approximately 3600 mm with the drier periods typically occurring between January and April (Brown et al. 1983). Mean monthly air temperatures are relatively constant throughout the year with lows of 21C in December-January, and highs of 24C between July-September. These forest types are described in greater detail in Brown et al. (1983), Scatena (1989) and Zou et al.(1995).

Plots were established in two separate catchments - one catchment was located at El Verde adjacent to the La Prieta stream (350-430 m elevation) and the other at Bisley on the # 3 gaged watershed (230-455 m elevation). Within each catchment three topographic areas were selected: (1) stream pools, (2) riparian zones and (3) the upslope area. The riparian zone is an area adjacent to the stream that is periodically flooded and characterized by mottling in the soil. During year 1 and 2, five replicate plots (10x10 m) were established within each topographic area of each catchment, resulting in 15 plots in each catchment. During year 5, only three replicate plots (15x30 m) were established in each topographic area of each catchment, resulting in 9 plots per catchment. The riparian plot was located immediately adjacent to the stream plot and the upslope plot was either immediately adjacent to or within 30 m of the riparian plot. The upslope plot was not located on a ridge. The sampling area in the Bisley #3 catchment included a 200 m long stream reach while the area at El Verde's La Prieta included a 100 m long stream reach. The plots were systematically located within that 100-200 m reach of each stream with a minimum distance of 30 m between plots to ensure their adequate interspersion while conserving some randomness (Krebs 1989). For example, the requisite number of deep stream pools were found along each selected stream reach and used as the stream plots. The riparian and upslope plots were then positioned perpendicular to the stream and near the stream plots. Plots were interspersed in each catchment to avoid spatial autocorrelation. Litterfall collectors were then randomly positioned within the plots. The plots varied somewhat in aspect and slope (Bisley had steeper slopes than El Verde) within each catchment.

Litterfall in the riparian and upslope plots was sampled using plastic laundry baskets (1,444 cm2 collection surface area - 29 cm deep and 38x38 cm wide at the open top) which were lined with fiberglass window screening (1 mm size openings). The baskets had large holes punched in the bottom to ensure that water drained from them. Plants that were 50 cm or taller were the only plants that contributed to the sampled litterfall. Litterfall occurring in the stream plots was sampled using a rectangular litterfall collector (30,000 cm2 collection surface area - 244 cm long, 122 cm wide and 61 cm deep), made of inert plastic mesh (1 mm size openings) and suspended ~ 1-2 m above the stream pool. During the first two years, two terrestrial litterfall collectors were placed in each of the five upslope and riparian plots for a total of 10 collectors at each of those topographic positions per catchment. However during the fifth year, five litterfall collectors were placed in each of the upslope and riparian plots but only three plots were established per topographic position per catchment for a total of 15 collectors at each topographic positions per catchment. For the stream plots only 1 stream litterfall collector was used in each of the five stream plots in each catchment during year 1 and 2 making a total of 5 stream litterfall collectors per catchment. However during the fifth year (as with the riparian and upslope plots) only 3 plots were established but 2 stream litterfall collectors per plot were erected making a total of 6 stream litterfall collectors per catchment.

Litterfall mass (g/m2) was recorded in these data files for specific dates (i.e., sampling time). However, for the summarization in the abstract these data were converted to monthly litterfall mass (g/m2) which were estimated by determining the number of days between a litterfall collection date (e.g., 'date=2') and the prior collection date (e.g., 'date=1'), divide the litterfall mass of 'date=2' by 'days since prior collection date' to calculate the 'estimated mass/day', then sum 'estimated daily mass' within each month for the 'estimated monthly mass'. The numerical manipulation and statistics were accomplished using Microsoft Excel for Windows v. 7.0a, Microsoft FoxPro v. 2.5 and The Sas System for Windows v. 6.11.

Boccheciamp, R.A. 1977. Soil survey of the Humacao area of eastern Puerto Rico. USDA Soil Cons. Serv., U.S. Gov. Print. Office, Washington, D.C.

Brown, S., A.E. Lugo, S. Silander, and L. Liegel. 1983. Research history and opportunities in the Luquillo Experimental Forest. U.S. Dep. Agric. For. Ser. Gen. Tech. Rep. SO-44. New Orleans, Louisiana.

Krebs, C.J. 1989. Ecological methodology. Harper and Row, New York.

Scatena, F.N. 1989. An introduction to the physiography and history of the Bisley Experimental Watersheds in the Luquillo Mountains of Puerto Rico. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. SO-72. New Orleans, Louisiana.

Zou, X., C.P. Zucca, R.B. Waide, and W.H. McDowell. 1995. Long- term influence of deforestation on tree species composition and litter dynamics of a tropical rain forest in Puerto Rico. Forest Ecol. and Management 78:147-157.

CROSS-REFERENCES (other data sets related to this one): LTERDBAS#51: MRCE/LTER Soil and Microbial Biomass Nitrogen; LTERDBAS #93: Litter decomposition in tabonuco forest before Hugo; LTERDBAS #94: Litterfall of the tabonuco forest before Hurricane Hugo; LTERDBAS #95: Litterfall along topographic gradients at lower Bisley; LTERDBAS #111: Litterfall in tabonuco (subtropical wet) forest in the Luquillo Experimental Forest, Puerto Rico (MRCE Litterfall data); LTERDBAS #115: Short-term disappearance of foliar litter of three tree species native to rain forest of Puerto Rico

SAMPLE LOCATION: Yale University, School of Forestry & Environmental Studies, Greeley Memorial Lab, 370 Prospect St., New Haven, CT 06511, USA, Basement Room 6

STORAGE SITES: Yale University, School of Forestry & Environmental Studies, Greeley Memorial Lab, 370 Prospect St., New Haven, CT 06511, USA, Room 212 computer room. Data on 3 1/2 floppy disk in storage container labeled Puerto Rico Data on shelf and on hard disk drive (C:\Puerto Rico) of Dell XPS computer. ITES, Data Manager's File DM-001 , Drawer #2

INVESTIGATOR'S ASSIGNED KEYWORDS: litter transfers, litterfall, Hurricane Hugo, Luquillo Experimental Forest, Puerto Rico, LUQ LTER site, stream, riparian and upslope areas, wood, leaf and miscellaneous tree tissues, tropical forest

LEF LTER OFFICIAL KEYWORDS (See table): BISLEY, EL VERDE, RIPARIAN, STREAM, TABONUCO, HURRICANE, LITTER FALL, SECOND FOREST, TREES, PEER REVIEWED JOURNAL

PUBLICATIONS:
Vogt, K.A., D.J. Vogt, P.Boon, A. Covich, F. N. Scatena, H. Asbjornsen, J.L. O'Hara, J. Pérez, T.G. Siccama, J. Bloomfield, J.F. Ranciato. 1997. Litter dynamics along stream, riparian and upslope areas following Hurricane Hugo, Luquillo Experimental Forest, Puerto Rico. Biotropica 28(4a):458-470.

DISSEMINATION: UNRESTRICTED

REASONS TO RESTRICT DATA IN THIS DATA SET BEYOND ITS TWO YEAR POLICY PERIOD*:

*WILL HAVE TO BE APPROVED BY LTER PRINCIPAL INVESTIGATORS: N. BROKAW, A. LUGO

SITES DESCRIPTIONS:

Geographical positional system (GPS) Coordinates for each location:

location	latitude	longitude
El Verde
La Prieta
Bisley #3

VARIABLES (ATTRIBUTES):

FILE NAME OR #ABOVE (all in which the variable appears)	All	All	All	All
ABBREVIATION (as it appears on the data file)	DATE	SITE	PLOT	LOCATION
NAME OF VARIABLE	Point in time	Location-site	Plot	Topographic position location
DEFINITION OF VARIABLE	Point in time (in mm/dd/yyyy) when litterfall collectors were sampled. No data is missing in this field.	Watershed code for research location-site. No data is missing in this field.	Code number of plot. No data is missing in this field.	Topographic position location. No data is missing in this field.
UNIT
PRECISION
RANGE OR LIST OF VALUES		1 = El Verde, 3 = La Prieta, 5 = Bisley #3	1 = stream pools; lowest downstrem, 2 = riparian zones; middle plot, 3 = the upslope area; upstream	1-16
DATA TYPE	datetime	integer	integer	integer
MISSING DATA CODES

VARIABLES:

FILE NAME OR #ABOVE (all in which the variable appears)	All	All	All	All
ABBREVIATION (as it appears on the data file)	REPLICATE	LEAVES	WOOD	RACHIS
NAME OF VARIABLE	Replicate plots	Biomass of leaves	Biomass of woody plant materials	Biomass of palm rachis
DEFINITION OF VARIABLE	Replicate plots (10x10 m) that were established within each topographic area of each catchment, resulting in 15 plots in each catchment during year 1 and 2 and during year 5, three replicate plots (15x30 m) that were established in each topographic area of each catchment, resulting in 9 plots per catchment. No data is missing in htis field.	Biomass of leaves dried at 70C (g/m2) (grams of litterfall leaves divided by the area of the litterfall). A bkank appears when data was not collected.	Biomass of woody plant materials < 1 cm (grams of litterfall wood divided by the area of the litterfall). A bkank appears when data was not collected.	Biomass of palm rachis (central stalk of the compund leaf of the palm) (grams of litterfall rachis, divided by the area of the litterfall). A bkank appears when data was not collected.
UNIT		gramsPerSquareMeter	gramsPerSquareMeter	gramsPerSquareMeter
PRECISION		0.01	0.00	0.01
RANGE OR LIST OF VALUES	1, 2, 3, 4, 5
DATA TYPE		numeric	numeric	numeric
MISSING DATA CODES

VARIABLES:

FILE NAME OR #ABOVE (all in which the variable appears)	All	All
ABBREVIATION (as it appears on the data file)	OTHER	TOTAL
NAME OF VARIABLE	Biomass of others litterfall materials	Biomass of total litterfall materials
DEFINITION OF VARIABLE	Biomass of others (e.g. seeds, fruits, flowers, non-identifiable, etc.) (grams of other litterfall divided by the area of the litterfall). A bkank appears when data was not collected.	Biomass of total (sum of leaves, wood, rachis, and others). A blank appears when data was not collected.
UNIT	gramsPerSquareMeter	gramsPerSquareMeter
PRECISION	0.01	0.01
RANGE OR LIST OF VALUES
DATA TYPE	numeric	numeric
MISSING DATA CODES

COMPUTATIONAL METHODS:

Variable Name	Formula
Wood	Collectors - dry mass of litterfall wood (g) / area of collector (m2) = litterfall mass (g) / m2
Ranchis	Collectors - dry mass of litterfall ranchis (g) / area of collector (m2) = litterfall mass (g) / m2
Other	Collectors - dry mass of other litterfall (g) / area of collector (m2) = litterfall mass (g) / m2
Total	Leaves+Wood+Rachis+Other

FOR DATA MANAGER USE ONLY

DATE OF LAST REVIEW: May 3, 2009
DATE OF LAST ENTRY: 1993
STAGE OF DATA SET MANAGEMENT (dates):
RECEIVED: 13 May 1997
CATALOGUED: 28 May 1997
ON-LINE: 28 May 1997
REVIEWED BY RESEARCHER: 13 May 1997
FILING MEDIA:
NAME OF DOCUMENTATION FILE: LTERDB95.FM*
NAME OF DATA FILE: abov8992.txt, abov992.txt
NAME OF ON - LINE CATALOG: LTERDBAS
RECORD #: 95
DOCUMENT TYPE: magnetic media only
PRIORITY TO BE ENTERED: N/A

Rev. date of this form: 18 March 2002