Hello everyone,
I am working for a few days already on a basic algorithm, very common in applied agronomy, that aims to determine the degree-days necessary for a given individual to reach a given growth stade. The algorithm (and context) is explained here: http://www.oardc.ohio-state.edu/gdd/glossary.htm , and so I implemented my function in R as follows: DD <- function(Tmin, Tmax, Tseuil, meanT, method = "DDsin") ### function that calculates the degree-days based on ### minimum and maximum recorded temperatures and the ### minimal threshold temperature (lower growth temperature) { ### method arcsin if(method == "DDsin"){ cond1 <- (Tmax <= Tseuil) cond2 <- (Tmin >= Tseuil) amp <- ((Tmax - Tmin) / 2) print((Tseuil-meanT)/amp) alpha <- asin((Tseuil - meanT) / amp) DD_ifelse3 <- ((1 / pi) * ((meanT - Tseuil) * ((pi/2) - alpha)) + amp*cos(alpha)) DD <- ifelse(cond1, 0, ifelse(cond2, (meanT - Tseuil), DD_ifelse3)) } ### method (Tmin + Tmax) / 2 else if(method == "DDt2"){ cond1 <- (meanT > Tseuil) DD <- ifelse(cond1,(meanT - Tseuil),0) } else{ stop("\nMethod name is invalid.\nMethods available = DDsin (sinus) or DDt2 (mean)\n") } return(DD) } BUT! When I try to process random data: library(reshape2) library(plyr) station <- rep(c("station1","station2","station3"), 20) values_min <- sample(-5:20, size = 60, replace = T) values_max <- sample(20:40, size = 60, replace = T) meanT <- ((values_min+values_max)/2) d <- data.frame(station,values_min,values_max,meanT) names(d) <- c("station", "values_min","values_max","meanT") x<-ddply(d, .(station), transform, t1 = cumsum(DD(values_min,values_max,0,meanT))) I get a warning on my alpha calculation (NaN produced); indeed, the values I give as argument to asin() are out of the range [-1:1], as the print() reveals. I can't figure out how to solve this issue, because the same algorithm works in Excel (visual basic). It is very annoying, especially because it seems that no occurence of such error using that algorithm can be found on Internet. Any help is welcome :) Thanks for your time P. |
On 12-03-19 7:42 AM, Letnichev wrote:
> Hello everyone, > > I am working for a few days already on a basic algorithm, very common in > applied agronomy, that aims to determine the degree-days necessary for a > given individual to reach a given growth stade. The algorithm (and context) > is explained here: http://www.oardc.ohio-state.edu/gdd/glossary.htm , and > so I implemented my function in R as follows: > > DD<- function(Tmin, Tmax, Tseuil, meanT, method = "DDsin") > ### function that calculates the degree-days based on > ### minimum and maximum recorded temperatures and the > ### minimal threshold temperature (lower growth temperature) > { > ### method arcsin > if(method == "DDsin"){ > cond1<- (Tmax<= Tseuil) > cond2<- (Tmin>= Tseuil) These look like useful diagnostics of out-of-range values, but you don't use them before the arcsin transformation. > amp<- ((Tmax - Tmin) / 2) > print((Tseuil-meanT)/amp) > alpha<- asin((Tseuil - meanT) / amp) > DD_ifelse3<- ((1 / pi) * ((meanT - Tseuil) * ((pi/2) - alpha)) + > amp*cos(alpha)) > > DD<- ifelse(cond1, 0, ifelse(cond2, (meanT - Tseuil), DD_ifelse3)) > } > > ### method (Tmin + Tmax) / 2 > else if(method == "DDt2"){ > cond1<- (meanT> Tseuil) > DD<- ifelse(cond1,(meanT - Tseuil),0) > } > > else{ > stop("\nMethod name is invalid.\nMethods available = DDsin (sinus) or DDt2 > (mean)\n") > } > return(DD) > } > > BUT! When I try to process random data: It's a good idea to use set.seed when trying to debug problems like this. Then you can construct a reproducible example. I'd also suggest getting rid of ddply at least for debugging; it makes it harder to see what's going on. Duncan Murdoch > > library(reshape2) > library(plyr) > > station<- rep(c("station1","station2","station3"), 20) > values_min<- sample(-5:20, size = 60, replace = T) > values_max<- sample(20:40, size = 60, replace = T) > meanT<- ((values_min+values_max)/2) > d<- data.frame(station,values_min,values_max,meanT) > names(d)<- c("station", "values_min","values_max","meanT") > > x<-ddply(d, .(station), transform, t1 = > cumsum(DD(values_min,values_max,0,meanT))) > > I get a warning on my alpha calculation (NaN produced); indeed, the values I > give as argument to asin() are out of the range [-1:1], as the print() > reveals. I can't figure out how to solve this issue, because the same > algorithm works in Excel (visual basic). > It is very annoying, especially because it seems that no occurence of such > error using that algorithm can be found on Internet. > Any help is welcome :) Thanks for your time > > P. > > -- > View this message in context: http://r.789695.n4.nabble.com/Issue-with-asin-tp4484462p4484462.html > Sent from the R help mailing list archive at Nabble.com. > > ______________________________________________ > [hidden email] mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by Letnichev
On 19-03-2012, at 12:42, Letnichev wrote: > Hello everyone, > > I am working for a few days already on a basic algorithm, very common in > applied agronomy, that aims to determine the degree-days necessary for a > given individual to reach a given growth stade. The algorithm (and context) > is explained here: http://www.oardc.ohio-state.edu/gdd/glossary.htm , and > so I implemented my function in R as follows: > > DD <- function(Tmin, Tmax, Tseuil, meanT, method = "DDsin") > ### function that calculates the degree-days based on > ### minimum and maximum recorded temperatures and the > ### minimal threshold temperature (lower growth temperature) > { > ### method arcsin > if(method == "DDsin"){ > cond1 <- (Tmax <= Tseuil) > cond2 <- (Tmin >= Tseuil) > amp <- ((Tmax - Tmin) / 2) > print((Tseuil-meanT)/amp) > alpha <- asin((Tseuil - meanT) / amp) > DD_ifelse3 <- ((1 / pi) * ((meanT - Tseuil) * ((pi/2) - alpha)) + > amp*cos(alpha)) > > DD <- ifelse(cond1, 0, ifelse(cond2, (meanT - Tseuil), DD_ifelse3)) > } > > ### method (Tmin + Tmax) / 2 > else if(method == "DDt2"){ > cond1 <- (meanT > Tseuil) > DD <- ifelse(cond1,(meanT - Tseuil),0) > } > > else{ > stop("\nMethod name is invalid.\nMethods available = DDsin (sinus) or DDt2 > (mean)\n") > } > return(DD) > } > > BUT! When I try to process random data: > > library(reshape2) > library(plyr) > > station <- rep(c("station1","station2","station3"), 20) > values_min <- sample(-5:20, size = 60, replace = T) > values_max <- sample(20:40, size = 60, replace = T) > meanT <- ((values_min+values_max)/2) > d <- data.frame(station,values_min,values_max,meanT) > names(d) <- c("station", "values_min","values_max","meanT") > > x<-ddply(d, .(station), transform, t1 = > cumsum(DD(values_min,values_max,0,meanT))) > > I get a warning on my alpha calculation (NaN produced); indeed, the values I > give as argument to asin() are out of the range [-1:1], as the print() > reveals. I can't figure out how to solve this issue, because the same > algorithm works in Excel (visual basic). That doesn't mean that Excel and/or Visual Basic gives correct answers. With the same input? Then what does Excel say that asin(-7.4) evaluates to? I tried asin(-1.2) and asin(-7.4) in LibreOffice Calc (3.5.0) and got #VALUE! (Error: wrong data type) twice. You'll have to present correct input to asin() if you want to avoid the NaN's. Berend > It is very annoying, especially because it seems that no occurence of such > error using that algorithm can be found on Internet. > Any help is welcome :) Thanks for your time > > P. > > -- > View this message in context: http://r.789695.n4.nabble.com/Issue-with-asin-tp4484462p4484462.html > Sent from the R help mailing list archive at Nabble.com. > > ______________________________________________ > [hidden email] mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by Letnichev
Hi,
You're not following the algorithm as given. The asin step shouldn't be done for all values, but only for the ones that don't meet the previous conditions. You're trying to calculate that step for ALL values, then only use certain ones. You must instead subset the values, THEN calculate that step. I would guess that your "working" Excel version does follow the correct algorithm, but it's hard to know for certain. Here's a version that more closely follows the given reference: MaxDailyTemp <- values_max MinDailyTemp <- values_min k <- 0 GDD <- rep(0, length(Tmin)) AvgDailyTemp <- (MaxDailyTemp + MinDailyTemp)/2 # if MaxDailyTemp < k # GDD = GDD + 0 # - add 0 # if MaxDailyTemp > k & MinDailyTemp > k # GDD = GDD + AvgDailyTemp - k GDD[MaxDailyTemp > k & MinDailyTemp > k] <- AvgDailyTemp[MaxDailyTemp > k & MinDailyTemp > k] - k # if MaxDailyTemp > k & MinDailyTemp < k # GDD = GDD + (1/pi) * [ (AvgDailyTemp – k) * ( ( pi/2 ) – arcsine( theta ) ) + ( a * cos( arcsine( theta ) ) ) ] a <- (MaxDailyTemp - MinDailyTemp)/2 theta <- ((k - AvgDailyTemp)/a) GDD[MaxDailyTemp > k & MinDailyTemp < k] <- (1/pi) * ( (AvgDailyTemp[MaxDailyTemp > k & MinDailyTemp < k] - k) * ( ( pi/2 ) - asin( theta[MaxDailyTemp > k & MinDailyTemp < k] ) ) + ( a[MaxDailyTemp > k & MinDailyTemp < k] * cos( asin( theta[MaxDailyTemp > k & MinDailyTemp < k] ) ) ) ) sum(GDD) Sarah On Mon, Mar 19, 2012 at 7:42 AM, Letnichev <[hidden email]> wrote: > Hello everyone, > > I am working for a few days already on a basic algorithm, very common in > applied agronomy, that aims to determine the degree-days necessary for a > given individual to reach a given growth stade. The algorithm (and context) > is explained here: http://www.oardc.ohio-state.edu/gdd/glossary.htm , and > so I implemented my function in R as follows: > > DD <- function(Tmin, Tmax, Tseuil, meanT, method = "DDsin") > ### function that calculates the degree-days based on > ### minimum and maximum recorded temperatures and the > ### minimal threshold temperature (lower growth temperature) > { > ### method arcsin > if(method == "DDsin"){ > cond1 <- (Tmax <= Tseuil) > cond2 <- (Tmin >= Tseuil) > amp <- ((Tmax - Tmin) / 2) > print((Tseuil-meanT)/amp) > alpha <- asin((Tseuil - meanT) / amp) > DD_ifelse3 <- ((1 / pi) * ((meanT - Tseuil) * ((pi/2) - alpha)) + > amp*cos(alpha)) > > DD <- ifelse(cond1, 0, ifelse(cond2, (meanT - Tseuil), DD_ifelse3)) > } > > ### method (Tmin + Tmax) / 2 > else if(method == "DDt2"){ > cond1 <- (meanT > Tseuil) > DD <- ifelse(cond1,(meanT - Tseuil),0) > } > > else{ > stop("\nMethod name is invalid.\nMethods available = DDsin (sinus) or DDt2 > (mean)\n") > } > return(DD) > } > > BUT! When I try to process random data: > > library(reshape2) > library(plyr) > > station <- rep(c("station1","station2","station3"), 20) > values_min <- sample(-5:20, size = 60, replace = T) > values_max <- sample(20:40, size = 60, replace = T) > meanT <- ((values_min+values_max)/2) > d <- data.frame(station,values_min,values_max,meanT) > names(d) <- c("station", "values_min","values_max","meanT") > > x<-ddply(d, .(station), transform, t1 = > cumsum(DD(values_min,values_max,0,meanT))) > > I get a warning on my alpha calculation (NaN produced); indeed, the values I > give as argument to asin() are out of the range [-1:1], as the print() > reveals. I can't figure out how to solve this issue, because the same > algorithm works in Excel (visual basic). > It is very annoying, especially because it seems that no occurence of such > error using that algorithm can be found on Internet. > Any help is welcome :) Thanks for your time > > P. > -- Sarah Goslee http://www.functionaldiversity.org ______________________________________________ [hidden email] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by Berend Hasselman
Yes, with the same input I had two different outputs with Excel and R. When printing a debug report of Excel, it showed no anomalies and I am certain it didn't calculate odd values (such as NaNs).
The way I coded was wrong, as Sarah said, I didn't follow completely the algorithm. The solution she suggested works perfectly, so I am out of trouble (for now :p ). Thanks for your time, P. |
In reply to this post by Sarah Goslee
Hello,
you're totally right, I tried first to control the flow with "if (MaxDailyTemp > k & MinDailyTemp < k){statement}" but it was a bit messy. Then ifelse() was supposed to help me out, but it didn't. Thank you for your time, your code works exactly as I want :) P. |
Free forum by Nabble | Edit this page |