Marta Sestelo | Universidade de Vigo (original) (raw)

Uploads

Papers by Marta Sestelo

Journal of Statistical Software

We present the R npregfast package via some applications involved with the study of living organi... more We present the R npregfast package via some applications involved with the study of living organisms. The package implements nonparametric estimation procedures in regression models with or without factor-by-curve interactions. The main feature of the package is its ability to perform inference regarding these models. Namely, the implementation of different procedures to test features of the estimated regression curves: on the one hand, the comparisons between curves which may vary across groups defined by levels of a categorical variable or factor; on the other hand, the comparisons of some critical points of the curve (e.g., maxima, minima or inflection points), studying for this purpose the derivatives of the curve.

The R Journal

One major goal in clinical applications of time-to-event data is the estimation of survival with ... more One major goal in clinical applications of time-to-event data is the estimation of survival with censored data. The usual nonparametric estimator of the survival function is the time-honored Kaplan-Meier product-limit estimator. Though this estimator has been implemented in several R packages, the development of the condSURV R package has been motivated by recent contributions that allow the estimation of the survival function for ordered multivariate failure time data. The condSURV package provides three different approaches all based on the Kaplan-Meier estimator. In one of these approaches these quantities are estimated conditionally on current or past covariate measures. Illustration of the software usage is included using real data.

Journal of the Royal Statistical Society: Series C (Applied Statistics)

In the nonparametric regression framework, issues of interest include the so-called factor-by-cur... more In the nonparametric regression framework, issues of interest include the so-called factor-by-curve interaction, where the effect of a continuous co-variate on response varies across groups defined by levels of a categorical variable. This study introduces a software application for R which per-forms inference in a nonparametric regression model. It describes the capa-bilities of the program for estimating these models (and their derivatives), for drawing different regression curves by factor levels and for drawing inferences about critical points (for instance, maxima or change points). Bootstrap methods were implemented to draw inferences from the deriva-tive curves, and binning techniques were applied to speed up computation in the estimation and testing processes. The software is illustrated using biological data.

ABSTRACT Analysis of such studies can be successfully performed using nonparametric regression mo... more ABSTRACT Analysis of such studies can be successfully performed using nonparametric regression models. In the nonparametric regression framework, issues of interest include the so-called factor-by-curve interaction, where the effect of a continuous covariate on response varies across groups defined by levels of a categorical variable. This study sought to compare regression curves and their derivatives that may vary across groups defined by different experimental conditions. For this purpose, we propose the use of local linear kernel smoothers. This study introduces a software application for R which performs inference in a nonparametric regression model. It describes the capabilities of the program for estimating these models (and their derivatives) and for drawing different regression curves by factor levels. The main feature of the package is its ability to draw inferences about critical points, such as maxima or change points linked to the derivative curves. Bootstrap methods were implemented to draw inferences from the derivative curves, and binning techniques were applied to speed up computation in the estimation and testing processes. The software is illustrated using biological data.

The aim of the present study was to estimate the length-weight relationship of Ruditapes decussat... more The aim of the present study was to estimate the length-weight relationship of Ruditapes decussatus and Ruditapes phillippinarum. For this purpose, this study was undertaken using data drawn from two estuaries along the coast of Cantabria (N Spain). The length-weight relationship of both species was estimated for both estuaries, using two models: a classic allometric model and a nonparametric model using local linear kernel smoothers. Additionally, derivatives were used for estimating a minimum size of capture for this species, corresponding to the size where the first derivative reaches the maximum. The confidence intervals, used to draw inferences from these derivatives curves, were constructed using bootstrap methods, and binning techniques were applied to speed up computation in the estimation and testing processes. Within this context, the models application and the individuals’ weight gain showed that (a) the nonparametric model resulted in a better fit of data for both specie...

Ocean & Coastal Management, 2013

The present study analyzes the weight gain patterns per unit of size and estimates the minimum su... more The present study analyzes the weight gain patterns per unit of size and estimates the minimum suitable catch size of two clam species: the carpet shell clam Ruditapes decussatus and the Manila clam Ruditapes philippinarum. For this purpose, data from the two largest estuaries along the northern coast of Spain (Cantabrian Sea) were used. The lengtheweight relationship of both studied species was estimated using two models: a classic allometric model and a nonparametric regression model based on local linear kernel smoothers. Additionally, first derivatives were used to estimate a minimum capture size for this species, corresponding to the size at which the first derivative reached the maximum. Within this context, the models application showed (a) the nonparametric model resulted in a better fit of data for both species (b) different minimum catch sizes for each species based on maximum length (49.5 mm for R. decussatus and 44.7 mm for R. philippinarum), both considerably larger than the currently established in EU and (c) an effect of estuaries and zones on individuals weight gain patterns. This confirmed the nonparametric model as an alternative approach to analyze the lengtheweight relationship for the studied species and to estimate a minimum suitable catch size of capture. The application of the specific catch sizes obtained in this study could lead to an increase in available commercial stocks of both species and positive effects on the conservation of the native species. In addition, the estimated zone-specific weight-gain patterns and minimum catch sizes could support a zone-based management. Therefore, this study provides a preliminary study and a starting point to consider the revision of the minimum legal size of the studied species in order to improve the current shellfishery management model.

An adequate nonparametric regression model is able to record specific patterns in the data that c... more An adequate nonparametric regression model is able to record specific patterns in the data that cannot be detected by a parametric model. In addition, quantile regression can provide a more complete description of functional changes than an exclusive focus on the least square regression. This chapter assesses the adequacy of a variety of nonparametric models to analyze the growth patterns of sea urchins by means of the length-weight relationship. For this purpose, data from fishery landings of green sea urchin Strongylocentrotus droebachiensis are used to analyze this relationship for lengths within the legal catch-size range (52.4 mm -76 mm) at two depths (i.e. shallow waters, 4.6 m, and deep waters, 7.6 m). Overall, this gives insight into the study of the minimum capture size. We apply a Kernel nonparametric regression model to determine both its suitability and applicability as an alternative to the classic allometric model, for the estimation of a minimum catch size directed to...

Ocean & Coastal Management, 2015

The suitability of a total-length-based, minimum capture-size and different protection regimes wa... more The suitability of a total-length-based, minimum capture-size and different protection regimes was investigated for the gooseneck barnacle Pollicipes pollicipes shellfishery in N Spain. For this analysis, individuals that were collected from 10 sites under different fishery protection regimes (permanently open, seasonally closed, and permanently closed) were used. First, we applied a non-parametric regression model to explore the relationship between the capitulum Rostro-Tergum (RT) size and the Total Length (TL). Important heteroskedastic disturbances were detected for this relationship, demonstrating a high variability of TL with respect to RT. This result substantiates the unsuitability of a TL-based minimum size by means of a mathematical model. Due to these disturbances, an alternative growthbased minimum capture size of 26.3 mm RT (23 mm RC) was estimated using the first derivative of a Kernel-based non-parametric regression model for the relationship between RT and dry weight. For this purpose, data from the permanently protected area were used to avoid bias due to the fishery. Second, the size-frequency distribution similarity was computed using a MDS analysis for the studied sites to evaluate the effectiveness of the protection regimes. The results of this analysis indicated a positive effect of the permanent protection, while the effect of the seasonal closure was not detected. This result needs to be interpreted with caution because the current harvesting based on a potentially unsuitable minimum capture size may dampen the efficacy of the seasonal protection regime.

Ocean & Coastal Management, 2015

This study was undertaken using data drawn from 5 sites along the Atlantic shoreline of Galicia (... more This study was undertaken using data drawn from 5 sites along the Atlantic shoreline of Galicia (Northwest Spain) for a period of 2 y. The length–weight relationship of Pollicipes pollicipes (Gmelin, 1789) was estimated to observe the way in which individuals of this species gain weight as they increase in size. A classic allometric model was used for the purpose. As an alternative, a more general nonparametric model was also estimated, using local linear kernel smoothers. Comparison of these two models showed that use of the nonparametric model resulted in a better fit of the data. In addition, derivatives were used for estimating a size of capture for this species. For the same purpose, we also estimated this crustacean’s mean size at sexual maturation (L50) and the number of broods that it spawns per annum. Individuals! weight gain, a female maturity size of 15.7 mm, and P. pollicipes! estimated 1.73 broods per annum tend to suggest a size of capture based on a rostrocarinal length of 21.50 mm.

articles by Marta Sestelo

The present study analyzes the weight gain patterns per unit of size and estimates the minimum su... more The present study analyzes the weight gain patterns per unit of size and estimates the minimum suitable catch size of two clam species: the carpet shell clam Ruditapes decussatus and the Manila clam Ruditapes philippinarum. For this purpose, data from the two largest estuaries along the northern coast of Spain (Cantabrian Sea) were used. The length–weight relationship of both studied species was estimated using two models: a classic allometric model and a nonparametric regression model based on local linear kernel smoothers. Additionally, first derivatives were used to estimate a minimum capture size for this species, corresponding to the size at which the first derivative reached the maximum. Within this context, the models application showed (a) the nonparametric model resulted in a better fit of data for both species (b) different minimum catch sizes for each species based on maximum length (49.5 mm for R. decussatus and 44.7 mm for R. philippinarum), both considerably larger than the currently established in EU and (c) an effect of estuaries and zones on individuals weight gain patterns. This confirmed the nonparametric model as an alternative approach to analyze the length–weight relationship for the studied species and to estimate a minimum suitable catch size of capture. The application of the specific catch sizes obtained in this study could lead to an increase in available commercial stocks of both species and positive effects on the conservation of the native species. In addition, the estimated zone-specific weight-gain patterns and minimum catch sizes could support a zone-based management. Therefore, this study provides a preliminary study and a starting point to consider the revision of the minimum legal size of the studied species in order to improve the current shellfishery management model.► Suitability of a kernel-based nonparametric model to analyze the length–weight relationship of clams. ► Estimation of minimum catch sizes of clams using the first derivative of the model. ► The estimated sizes were higher than the currently established ones for both species.

A mathematical for the detection of SO2 emission episodes was developed.A generalized additive mo... more A mathematical for the detection of SO2 emission episodes was developed.A generalized additive model and an algorithm for variable selection were used.SO2 concentrations and meteorological variables were considered.The best prediction is reached with only two terms of the time series.Meteorological variables were found not to be significant covariates.The aim of this paper is to predict time series of SO2 concentrations emitted by coal-fired power stations in order to estimate in advance emission episodes and analyze the influence of some meteorological variables in the prediction. An emission episode is said to occur when the series of bi-hourly means of SO2 is greater than a specific level. For coal-fired power stations it is essential to predict emission episodes sufficiently in advance so appropriate preventive measures can be taken. We proposed a methodology to predict SO2 emission episodes based on using an additive model and an algorithm for variable selection. The methodology was applied to the estimation of SO2 emissions registered in sampling locations near a coal-fired power station located in Northern Spain. The results obtained indicate a good performance of the model considering only two terms of the time series and that the inclusion of the meteorological variables in the model is not significant.

Reliable information on the geographic location of individual points using GPS (Global Positionin... more Reliable information on the geographic location of individual points using GPS (Global Positioning System) receivers requires an unobstructed line of sight from the points to a minimum of four satellites. This is often difficult to achieve in forest environments, as trunks, branches and leaves can block the GPS signal. Forest canopy can be characterized by means of dasymetric parameters such as tree density and biomass volume, but it is important to know which parameters in particular have a bearing on the accuracy of GPS measurements. We analyzed the relative influence of forest canopy and GPS-signal-related variables on the accuracy of the GPS observations using a methodology based on linear regression models and bootstrapping and compared the results to those for a classical variable-selection method based on hypothesis testing. The results reveal that our methodology reduces the number of significant variables by approximately 50% and that both forestry and GPS-signal-related variables are significant.

In multiple regression models, when there are a large number (p) of explanatory variables which m... more In multiple regression models, when there are a large number (p) of explanatory variables which may or may not be relevant for predicting the response, it is useful to be able to reduce the model. To this end, it is necessary to determine the best subset of q (q ≤ p) predictors which will establish the model with the best prediction capacity. FWDselect package introduces a new forward stepwisebased selection procedure to select the best model in different regression frameworks (parametric or nonparametric). The developed methodology, which can be equally applied to linear models, generalized linear models or generalized additive models, aims to introduce solutions to the following two topics: i) selection of the best combination of q variables by using a step-by-step method; and, perhaps, most importantly, ii) search for the number of covariates to be included in the model based on bootstrap resampling techniques. The software is illustrated using real and simulated data.

Journal of Statistical Software

The R Journal

Journal of the Royal Statistical Society: Series C (Applied Statistics)

Ocean & Coastal Management, 2013

Ocean & Coastal Management, 2015

The present study analyzes the weight gain patterns per unit of size and estimates the minimum su... more The present study analyzes the weight gain patterns per unit of size and estimates the minimum suitable catch size of two clam species: the carpet shell clam Ruditapes decussatus and the Manila clam Ruditapes philippinarum. For this purpose, data from the two largest estuaries along the northern coast of Spain (Cantabrian Sea) were used. The length–weight relationship of both studied species was estimated using two models: a classic allometric model and a nonparametric regression model based on local linear kernel smoothers. Additionally, first derivatives were used to estimate a minimum capture size for this species, corresponding to the size at which the first derivative reached the maximum. Within this context, the models application showed (a) the nonparametric model resulted in a better fit of data for both species (b) different minimum catch sizes for each species based on maximum length (49.5 mm for R. decussatus and 44.7 mm for R. philippinarum), both considerably larger than the currently established in EU and (c) an effect of estuaries and zones on individuals weight gain patterns. This confirmed the nonparametric model as an alternative approach to analyze the length–weight relationship for the studied species and to estimate a minimum suitable catch size of capture. The application of the specific catch sizes obtained in this study could lead to an increase in available commercial stocks of both species and positive effects on the conservation of the native species. In addition, the estimated zone-specific weight-gain patterns and minimum catch sizes could support a zone-based management. Therefore, this study provides a preliminary study and a starting point to consider the revision of the minimum legal size of the studied species in order to improve the current shellfishery management model.► Suitability of a kernel-based nonparametric model to analyze the length–weight relationship of clams. ► Estimation of minimum catch sizes of clams using the first derivative of the model. ► The estimated sizes were higher than the currently established ones for both species.

Abstract We evaluated specific habitat features (bottom substrate type, depth, temperature, seaso... more Abstract We evaluated specific habitat features (bottom substrate type, depth, temperature, season and latitude-longitude) at random locations in the C{\'\i}es Islands (Galician Atlantic Islands National Park, Northwest Spain) to determine their impact on Sepia officinalis spawning habitat use. We performed underwater visual transects via scuba diving. In total, 94 transects were conducted between April 30, 2012 and July 17, 2015. Habitat features were evaluated as predictors of the presence/absence of egg clusters using Generalised Additive Models. Statistical analyses revealed the following: (a) the probability of finding eggs was significantly higher on hard bottom substrate type containing sea fans and sea tube worms; (b) the sectors where the egg presence was highest were the most shelters zones of Bajo de Vi{\~n}os and Piedra del Borr{\'o}n, two hard bottom shoals covered by sea fans and sea worms and located in the central C{\'\i}es Islands or Faro Island; (c) The preferential depth ranged from 8 to 13 m; (d) as temperatures rose, the probability of finding eggs decreased, which was a sign of the seasonality of the spawning event of the species; and (e) eggs were most common in the winter. A total of 18 additional non-random underwater visual transects were then performed in the Vi{\~n}os and Borr{\'o}n shoals between December 10, 2014 and August 20, 2015. We calculated the number of egg clusters found in two fenced areas (0.04 ha) located in Vi{\~n}os and Borr{\'o}n, which showed that Vi{\~n}os was a most suitable spawning habitat than Borr{\'o}n. The number of eggs per cluster varied from 4 to 82. The spawning period for this species lasted from the beginning of December until the end of June, when bottom temperature ranged from 12.5 ,circ\,^{\circ},circC to 14.75 ,circ\,^{\circ},circC. These microhabitat data reveal indicators of common cuttlefish spawning habitat utilisation and should help identify targets for habitat improvement projects and ecosystem management approaches within this national park and elsewhere.

Abstract The selection of spawning habitat of a population of Octopus vulgaris that is subject to... more Abstract The selection of spawning habitat of a population of Octopus vulgaris that is subject to a small-scale exploitation was studied in the C{\'\i}es Islands within the National Park of the Atlantic Islands of Galicia (NW Spain). The technique used was visual censuses by scuba diving. We conducted 93 visual censuses from April 2012 to April 2014. The total swept area was 123.69 ha. Habitat features (season, depth, zone, bottom temperature, swept area, bottom substrate type, and creels fishing impact) were evaluated as predictors of the presence/absence of spawning dens using \{GAM\} models. O. vulgaris has a noteworthy preference for spawning in areas with hard bottom substrate and moderate depth (approximately 20 m). The higher density of spawning dens (1.08 ha−1) was found in a surveyed area of 50.14 ha located in the northeastern part of the northern C{\'\i}es Island. We propose to protect the area comprised from Punta Escodelo to Punta Ferreiro between 5 and 30 m depth. This area has a surface of 158 ha equivalent to 5.98% of the total marine area of the C{\'\i}es islands. The strengths and weaknesses of a management strategy based on the protection of the species' spawning habitat are discussed.

Four visual censuses targeting Octopus vulgaris living in dens on sandy bottoms were carried out ... more Four visual censuses targeting Octopus vulgaris living in dens on sandy bottoms were carried out from June to October 2013 in the National Park of the Atlantic Galician Islands (NW Spain). Censuses were undertaken by scuba diving between 5 and 21 m depth in daytime. The total area swept was 13.75 ha. There were no significant differences between octopus presence in dens during open and closed fishing seasons. Depth had a significant negative relationship with occupancy. The average number of dens per 1000 m 2 was 3.84±0.84 in June and 3.89 in October. The area per den was 260 m 2 . Den number estimations varied between 1586 and 2057. The largest number of dens (76.5%) was found between 5 and 10 m depth. Den distribution was clumped. No significant differences were found between octopus size classes (small, medium and large) and den diameter. Associate dens were observed. There were no significant differences in den diameter and shell types found around the middens. Many dens could be "permanent". Drilling bivalve shell behaviour is discussed. The surveyed area had around 1100 individuals, mainly small specimens. No significant differences were found between octopus size and depth. Substrate, den type and food abundance and availability (especially razors Ensis arcuatus) seem to be the main factors influencing dens and octopus density and distribution. Den availability does not appear to be a limiting factor in this case. Temperature, den availability, predators and fishing pressure influencing density and distribution are discussed. Rodas inlet may be a preferential habitat for O. vulgaris individuals ranging from 200 to 2000 g, but especially small specimens (≤1000 g).