Long-TermData from Fields Recovering after Sugarcane, Banana, and Pasture Cultivation in Ecuador

I report here on an ongoing permanent plot study in areas recovering from agriculture in Ecuador. ese plots were set up in 1995 at Maquipucuna Reserve where the forest is tropical lower montane. e study consists of replicate �elds in three past crop types (Sugarcane, �anana, and Pasture) for a total of six �elds. Each �eld was �rst divided into 25 continuous 2m × 5m subplots which together form a 10m × 25m plot with the longest side bordering the adjacent forest. en starting in 1996, and continuing annually every year since, each subplot has been sampled for percent cover of all plants and diameter at breast height (dbh) for all trees whose dbh is greater than or equal to 1 cm. I have used that data in these published studies: (1) species composition and life form, richness, and basal area trends, (2) computation of all positive and negative pairwise species associations, (3) relationships between richness and productivity over time, (4) dominance-diversity curves, and (5) de�nition and quanti�cation of old �eld plant communities. Finally with the help of the LTER program in Puerto Rico, this sampling continues, with 2012 marking the sixteenth year of continuous annual sampling.


Introduction
Conducted for decades over many of the most common gradients on earth, studies of succession aer agriculture-also called postagricultural or old �eld succession-have helped ecologists gain insights into the processes that structure plant communities and into the role of history and initial conditions in community development [1,2].Examination of old �elds has led to the rise of important ecological theories, such as the initial �oristic composition hypothesis [3] and the resource ratio hypothesis [4].Using permanent plots [5] established at abandonment as the backbone of old �eld research (e.g., the �uell-Small old �eld plot study in New Jersey, USA [6,7]), scientists have been successful in �nding the pattern of response over time aer abandonment from crops [8] and in discovering many of the mechanisms that determine the pattern [9,10].Furthermore, studies of postagricultural succession have provided a framework for comparison and evaluation of various theories of succession and community development [11].
e nature of the recovery of old �elds depends on their disturbance regime, which is characterized by severity, size, spatial location, and frequency [38].e measurement of severity may be best accomplished by sampling the loss of plant biomass and determining whether such loss includes loss of plant meristems [4,39,40].Severity of old �elds is moderate, compared to the more severe landslides [41,42] and the less severe treefall [43] because old �elds maintain an intact soil pro�le, roots, and humus layer despite the removal of above-ground vegetation.In size, old �elds do not usually exceed a few tens of hectares (they are larger, however, than many Neotropic treefall gaps [44]), and crops that are planted in them oen require speci�c locations (e.g., Coffee grows well in the mountains while Banana and Sugarcane do best in the valleys).
Frequency of old �eld disturbance includes the number of times a �eld has been cultivated, the duration of each cropping period, the time between fallow periods, and the order of crop rotation.�sually �elds are fallow longer than they are in crop and can be either reused with different crops and fallow periods until abandoned or used for 5-10 years until productivity declines [45].As the length of time an area is in crop increases, the length of successional, or recovery, time may also lengthen.For example, a short cropping period can lead to large contributions from the seed bank, and trees may regenerate quickly from stump or root sprouts whereas an increased cropping period brings about the decay of rootstocks and tree stumps, thus providing for grass invasion and prolonged tree invasion.Among other conditions that in�uence old �eld recovery are historical effects of the past crop (a crop's signature [1,2,46]), the year of abandonment, the season of abandonment, the percent of border with forest [13], and a �eld's plowing record (whether the �eld was plowed under or le fallow when abandoned [7]).
Studies of postagricultural succession in the temperate zone have demonstrated the particular importance of past crops in determining old �eld patterns (e.g., alteration of successional pathways and species composition and abundance for up to 8 years aer abandonment [1,2,46]).Hence, I focused in this Neotropical study on the examination of �elds aer common Neotropic crops, such as Sugarcane (Saccharum officinarum), Banana (Musa spp.), and Coffee (Coffea spp.(see Puerto Rico studies reviewed in [38]), and aer pasture grasses that usually establish themselves aer cropping but can also be purposely planted (e.g., Setaria sphacelata).Pastures are a kind of old �eld succession for several reasons.First, they are usually converted from depleted agricultural �elds when cows are allowed to enter.Second, they may continue to have remnants of the previous past crops growing in them for some years.ird, they can be planted with exotic grasses or be colonized by local grass species due to cattle activity (e.g., trampling of vegetation, creation of hummocks, deposition of dung, and soil compaction).
For the studies reported here, all �elds were �rst cleared from forest using shiing or "slash and burn" [47] agriculture, which occurs when natural vegetation is �rst cut down to the ground and then burnt.Although burning produces a short pulse of nutrients like phosphorus, it generally volatilizes both organic matter and nitrogen while raising soil temperature and, in the long term, makes soil poor in both nutrients and the seed bank.Shiing agriculture also entails a rotation of �elds rather than crops that is accompanied by long fallow periods.e decision as to when to rotate �elds and for how long to leave them fallow depends on such factors as weed and insect loads in the �elds, soil fertility, production, labor availability, and local dietary needs [45].is forestclearing method causes carbon loss, a mixing of the soil horizons, and a higher degree of soil aeration [10].Shiing agriculture also lowers the level of advanced regeneration from sapling banks and from root/stem resprouting so that seed and seedling dynamics [48] dominate regeneration aer agriculture.
�ld �eld recovery goes through several structural and conceptual stages: (1) domination by past crop and its various effects, (2) patch dynamics of grass, past crop remnants, asexual shrubs, and other plants, (3) tree invasion, and (4) development of a closed-canopy secondary forest.Aer a closed canopy has developed, old �eld plants may eventually enter a thinning state when dead plants in the canopy are replaced not with new individuals but with modi�ed growth of existing plants.ere is a physiological limit to this kind of growth, however, and at some point every plant will die and be replaced.Indeed plants are always involved in the on-going dynamic process of plant-plant replacement [49,50].Plantplant replacement does not necessarily, however, involve two plants only for more than one plant can be replaced by just one or only one plant can be replaced by more than one.As plant-plant replacements proceed, the two spaces each plant has (its phytospace de�ned by its biomass and necromass, and its surrounding neighborhood space) can be occupied and reoccupied.Because plant replacements occur slowly, permanent plots are needed in order to sample the dynamic replacements all plants are engaged in.
Replacements create changes in plant abundances, which may also create changes in plant composition that lead to the emergence of other patterns at larger spatial and temporal scales (e.g., successional rate and direction, patch dynamics, plant distribution, exotic plant invasion, and plant associations (Tables 1-6)).Care must be taken, however, with the observation of these plant patterns.For example, even though individual plants continue to be replaced, preservation of species abundances and/or species composition at a larger scale of organization may still occur, giving the impression of a coexistence of species (when the identity or the number of plant species remains the same despite changes in plant abundances) or of a plant community at equilibrium (when neither plant species nor plant abundances change).
Existence of high-level patterns that do not change very much over time should not obscure the fact that such patterns are the product of a dynamic and ever-changing process of plantplant replacement.Any plant pattern is only a snapshot of the vegetation at a given instance in time and at a speci�c scale of observation whereas plant-plant replacements occur continuously.
Replacements are controlled by species tolerances and mechanisms which affect plant growth and the de�nition of plant spaces.Tolerances include germination and growth while mechanisms encompass dispersal, seed pool, seed predation, seed pathogens, seed germination, seedling predation, seedling pathogens, seedling herbivory, and/or seedling competition [51][52][53][54][55]. e process of plant-plant replacement is at the heart of the dynamic nature of terrestrial plant communities both over successional time and over space  (e.g., at the edges of plant distributions where gradients have a major effect on plants [56]).It seems that plants wait not to be counted (sensu [57]) but to be replaced!us the ultimate reality of plant communities is the plant-plant replacements, for they produce vegetation patterns at all larger temporal and spatial scales.

Methodology
e study site comprises recovering Banana plantations, Sugarcane plantations, and seeded pastures at the Maquipucuna Reserve, Ecuador (0 ∘ 05 � N, 78 ∘ 37 � W; http://www.maqui.org/[53,54,[58][59][60][61]. Maquipucuna lies between 1200 m and 1800 m and is classi�ed as tropical lower montane wet forest [62].It has deeply dissected drainages with steep slopes and receives between 2 m and 5 m of rainfall a year.e temperature ranges between 14 ∘ C and 25 ∘ C, with an average temperature of 18 ∘ C. e reserve's fertile Andisol soil is developed from volcanic ash deposits [42]. In June of 1996, six just abandoned agricultural �elds were selected for study: two recent Sugarcane (S. officinarum) plantations denoted le and right when facing the forest (Tables 3 and 4), two recent Banana (Musa sp.) plantations denoted le and right when facing the forest (Tables 1 and 2), and two recent pastures seeded in S. sphacelata denoted le and right when facing the forest (Tables 5 and 6).Although native to Africa, this Setaria grass is being planted more and more extensively in the Neotropics because it can withstand heavy grazing by spreading asexually and can form large crowns.Well accepted by cattle, the Setaria grass is tolerant of relatively cold temperatures and short-term waterlogged conditions, both common at the Maquipucuna Reserve.In each �eld, twenty-�ve 5 m × 2 m contiguous plots were laid out (Figure 1).All six �elds were 250 m 2 rectangles located (1) within a few hundred meters of each other, (2) at the lower elevations of the reserve, and (3) with the 25 m plot border next to primary forest in order to facilitate comparison with the Buell-Small plot study in New Jersey, USA, which had the same design [7].e plots did not have any remnant trees or sprouting tree roots at the beginning of the study, and their tree seed bank was very small [55].
Starting in 1997, each of the 5 m × 2 m subplots of each of the six plots was sampled annually to identify each of these plant species and (1) for percent cover of each  plant species-an indication of a species' ability to capture light and, therefore, to dominate these areas in the process of becoming forested communities-estimated visually in relation to each plot's area, and (2) for diameter at breast height (dbh) of each tree stem at least 1 cm.Maquipucuna plant taxonomists, trained at the University of Georgia, USA, where voucher specimens are kept on �le [42,61], assisted in the identi�cation of species by using specimens located on site.e 2012 sampling will mark the sixteenth year a�er abandonment for these �elds.I have published several analyses of this dataset [8,38,55,[63][64][65].

Dataset Description
e dataset associated with this Dataset Paper consists of 12 items which are described as follows.subplot within the plot in the study �eld (see Figure 1) and is an integer between 1 and 25.Species is the scienti�c name of species observed.Cover Percentage (%) is the percent cover for this species and is an integer between 0 and 100.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts, an observation is entered in this �eld.In the  1) and is an integer between 1 and �5.Species is the scienti�c name of species observed.Cover Percentage (%) is the percent cover for this species and is an integer between 0 and 100.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts, an observation is entered in this �eld.In the  1) and is an integer between 1 and �5.Species is the scienti�c name

Dataset Item 1 (Table
of species observed.Cover Percentage (%) is the percent cover for this species and is an integer between 0 and 100.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts, an observation is entered in this �eld.In the table, empty cell means missing value.).Ecuador old �eld�s percent cover �le: Pasture right �eld.It consists of �ve columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure 1) and is an integer between 1 and �5.Species is the scienti�c name of species observed.Cover Percentage (%) is the percent cover for this species and is an integer between 0 and 100.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts� an observation is entered in this �eld.�n the table� empty cell means missing value.).Ecuador old �eld�s percent cover �le� Sugarcane le� �eld.�t consists of �ve columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure 1) and is an integer between 1 and 25.Species is the scienti�c name of species observed.Cover Percentage (%) is the percent cover for this species and is an integer between 0 and 100.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts� an observation is entered in this �eld.�n the table� empty cell means missing value.).Ecuador old �eld�s percent cover �le� Sugarcane right �eld.�t consists of �ve columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure 1) and is an integer between 1 and 25.Species is the scienti�c name of species observed.Cover Percentage (%) is the percent cover for this species and is an integer between 0 and 100.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts� an observation is entered in this �eld.�n the table� empty cell means missing value.).Ecuador old �eld�s percent tree stem �le� Banana le� �eld.�t consists of seven columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure 1) and is an integer between 1 and 25.Tag Number is a unique number for each main stem in the entire LFDP.e main stem tag number is the number for an individual plant.No data is missing in this �eld.Species is the scienti�c name of species observed.Height is the height measured before cutting stem.Basal Diameter is the diameter at breast height of the tree and has a minimum of 1 cm.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts� an observation is entered in this �eld.�n the table� empty cell means missing value.).Ecuador old �eld�s percent tree stem �le� Banana right �eld.�t consists of seven columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure 1) and is an integer between 1 and 25.Tag Number is a unique number for each main stem in the entire LFDP.e main stem tag number is the number for an individual plant.No data is missing in this �eld.Species is the scienti�c name of species observed.Height is the height measured before cutting stem.Basal Diameter is the diameter at breast height of the tree and has a minimum of 1 cm.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts� an observation is entered in this �eld.�n the table� empty cell means missing value.).Ecuador old �eld�s percent tree stem �le� Pasture le� �eld.�t consists of seven columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure 1) and is an integer between 1 and 25.Tag Number is a unique number for each main stem in the entire LFDP.e main stem tag number is the number for an individual plant.No data is missing in this �eld.Species is the scienti�c name of species observed.Height is the height measured before cutting stem.
Basal Diameter is the diameter at breast height of the tree and has a minimum of 1 cm.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts, an observation is entered in this �eld.In the  1) and is an integer between 1 and 25.Tag Number is a unique number for each main stem in the entire LFDP.e main stem tag number is the number for an individual plant.No data is missing in this �eld.Species is the scienti�c name of species observed.Height is the height measured before cutting stem.Basal Diameter is the diameter at breast height of the tree and has a minimum of 1 cm.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts, an observation is entered in this �eld.In the  1) and is an integer between 1 and 25.Tag Number is a unique number for each main stem in the entire LFDP.e main stem tag number is the number for an individual plant.
No data is missing in this �eld.Species is the scienti�c name of species observed.Height is the height measured before cutting stem.Basal Diameter is the diameter at breast height of the tree and has a minimum of 1 cm.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts, an observation is entered in this �eld.In the  1) and is an integer between 1 and 25.Tag Number is a unique number for each main stem in the entire LFDP.e main stem tag number is the number for an individual plant.
No data is missing in this �eld.Species is the scienti�c name of species observed.Height is the height measured before cutting stem.Basal Diameter is the diameter at breast height of the tree and has a minimum of 1 cm.Comments is the observation made at the �eld about the living condition of the tree.When the tree is found dead for the �rst time or when it sprouts, an observation is entered in this �eld.In the table, empty cell means missing value.

Concluding Remarks
e establishment and regular sampling of permanent plots have proven invaluable in the investigation of areas recovering back to forest aer cultivation.In particular here in the Americas, such plots have been a major part of those investigations [9,38].With the assistance of the LTER program, I will continue that monitoring in �elds recovering from the common Neotropical crops of Sugarcane, Banana, and seeded pasture.

Dataset Availability
e dataset associated with this Dataset Paper is dedicated to the public domain using the CC0 waiver and is available at http://dx.doi.org/10.7167/2013/468973/dataset. In addition, all the data from the six plots are housed in the archives of the Luquillo (LUQ) Experimental Forest, which is a Long-Term Ecological Research (LTER) site, as LTERDATB#101.One may visit their website (http://luq.lternet.edu/)for access.

�on��ct of �nterests
ere is no con�ict of interests in the access or publication of this dataset.

F 1 :
Layout of each of the six 25 m × 10 m plots with each of its 5 m × 2 m subplots labeled as they are indicated in the datasets for each �eld.

T 6 :
Signi�cant Spearman rank correlation coe�cients among all plant species in the Ecuador Right Pasture over the �rst 10 years of succession.Signi�cant positive associations are indicated by a "+" and signi�cant negative associations are indicated by a "−".Plant species and families are indicated as follows: Setaria sphacelata (A), Erythrina megistophyla (B), and Polpodiaceae (C).
Dataset Item 2 (Table).Ecuador old �eld�s percent cover �le: Banana right �eld.It consists of �ve columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure table, empty cell means missing value.
DatasetItem 3 (Table).Ecuador old �eld�s percent cover �le: Pasture le� �eld.It consists of �ve columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure

Table ) .
table, empty cell means missing value.Ecuador old �eld�s percent tree stem �le: Pasture right �eld.It consists of seven columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure

Table ) .
table, empty cell means missing value.Ecuador old �eld�s percent tree stem �le: Sugarcane le �eld.It consists of seven columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure

Table ) .
table, empty cell means missing value Ecuador old �eld�s percent tree stem �le: Sugarcane right �eld.It consists of seven columns described as follows.e column Date is the date which the sampling took place.No data is missing for this �eld.Plot Number is the subplot within the plot in the study �eld (see Figure